Tuesday - Conference Calendar

Transcription

Philadelphia Back Matter_Philadelphia Back Matter 10/7/15 8:54 AM Page 258
TA01
INFORMS Philadelphia – 2015
Tuesday, 8:00am - 9:30am
between global infrastructure, epidemiology, economics, government policy, and
regional and/or international populations. This presentation summarizes the
development (web and desktop) and use of the GlobalCURE tool. In our analysis,
we specifically focus on the interplay of factors across levels of aggregation (e.g.,
tract through country).
■ TA01
2 - Optimization Planning Tool for Urban Search Missions
Daniel Faissol, Lawrence Livermore National Laboratory,
Livermore, CA, United States of America, faissol1@llnl.gov,
Claudio Santiago, Richard Wheeler, Thomas Edmunds
01-Room 301, Marriott
Military Manpower and Force Management
Sponsor: Military Applications
Sponsored Session
We present a prototype tool to support planning of radiological and nuclear
search missions in an urban environment using mobile detectors. Two distinct
problems are considered with proposed solutions: (1) a nonconvex optimization
problem that solves for detector dwell times and locations that maximize the
probability of detection for building interiors, and (2) a multiple vehicle routing
problem on a directed multigraph that solves for the maximum net benefit given
a fixed total search time.
Chair: Andrew Hall, COL, U.S. Army, 4760 40th St N, Arlington, VA,
United States of America, AndrewOscarH@aol.com
1 - Air Force Officer Accession Planning: Addressing Key Gaps in
Meeting Career Field Academic Degree Requirements
Tara Terry, Operations Researcher, RAND Corporation, 1200 S.
Hayes St., Arlington, VA, 22202, United States of America,
tterry@rand.org
3 - Optimal Sonar Deployment in a Maritime Environment:
A Fortification Approach
Taofeek Biobaku, University of Houston, Houston, TX,
United States of America, tobiobaku@uh.edu, Gino Lim,
Jaeyoung Cho, Hamid Parsaei, Seon Jin Kim
The goal of the Air Force officer accession process is to ensure the USAF accesses
officers with the knowledge, skills and attributes to perform missions in particular
career fields. Key to this goal for non-rated officers is establishing and enforcing
academic degree requirements. We uncovered gaps in accession processes that
undermine meeting career field education requirements. We introduced
recommendations toward correcting the accession process and meeting career
fields academic needs.
The safety and integrity of maritime assets continue to be of paramount
importance in world trade and economy. The marine-based trilevel problem
remains computationally challenging. The inherent challenges increase with the
risk analysis approach we adopt. We propose algorithms based on modifications of
Benders’ decomposition; and column-and- constraint general algorithms to
attempt an optimal solution. Thereafter, we compare solutions on these two
algorithms using a case study.
2 - A Methodology for Estimating Caseload in the U.S. Army’s
Disability Rating Process
James Broyles, Operations Researcher, RAND Corporation, 1776
Main Street, Santa Monica, CA, 90401, United States of America,
jbroyles@rand.org, Mustafa Oguz
4 - A Mothership-based UAV Routing Problem in Support of
Counterfire Operations
Jaeyoung Cho, University of Houston, 333 Dominion Dr.,
#1021, Katy, TX, 77450, United States of America,
uncmac.rokag@gmail.com, Taofeek Biobaku, Seon Jin Kim,
Gino Lim
As U.S. Army soldiers separate from service, a portion of them enter the disability
rating process to obtain a rating that determines their level of benefits and
compensation. The process involves several evaluation steps and appeal processes
that cause highly variable and sometimes long processing durations. This research
presents a methodology that uses a non-Markovian probability model for
estimating disability rating caseload given forecasted future soldier separations.
We describe a model for routing UAVs which are launched and recovered from
airborne drone carriers. We formulate and solve this problem with a given fleet of
UAVs subject to technical and operational constraints. The spatio-temporal model
captures important aspects of a UAV deployment in counterfire operations
including collaboration tactics and overlapping observation. The model is
designed to provide an insight into issues associated with operating UAVs aided
counterfire operations system.
3 - Aligning Officer Personnel Requirements with a Sustainable
Career Lifecycle Model
Michael Needham, DCS G-1, HQDA, 300 Army Pentagon,
Washington, DC, United States of America,
michael.p.needham2.mil@mail.mil
The U.S. Army is at a critical juncture in determining a supportable military
personnel structure that is limited by mandated force structures. Personnel
structure adjustments drive near-term force-shaping personnel policies, such as
accessions, promotions, and separations. We identify sustainable standards of
grade using historical data while accounting for future personnel management
policies. The model uses sixteen years of historical data as a foundation to
determine future behavior.
■ TA03
Scheduling in Practice
Cluster: Scheduling and Project Management
Invited Session
4 - Army Officer Grade Distribution for the Army
Competitive Category
Francisco Baez, DCS G-1, HQDA, 300 Army Pentagon,
Washington, DC, United States of America,
francisco.r.baez.mil@mail.mil
Chair: Emrah Cimren, Nike, 1 SW Bowerman Dr., Beaverton, OR,
97005, United States of America, Emrah.Cimren@nike.com
1 - A Sample-Gradient-Based Algorithm for Multiple-OR and PACU
Surgery Scheduling
Miao Bai, Lehigh University, 200 W Packer Ave, Bethlehem, PA,
18015, United States of America, mib411@lehigh.edu,
Gregory Tonkay, Robert Storer
The Army’s Grade Structure has become significantly senior impacting the
potential health of the current and future force by reducing selectivity and
competition rates, and forcing early promotions. The propose distribution of
officers focuses on re-balancing grade structure for each career management field
to ensure balance and health of the force by ensuring leader-to-led ratios, quality,
and viable career paths for all soldiers.
We address a multiple-OR surgery scheduling problem constrained by shared
PACU capacity within the block-booking framework. Given the surgery sequence,
a Discrete Event Dynamic System-based stochastic optimization model is
formulated in order to minimize the cost incurred by patient waiting time,
surgeon idle time, OR blocking time, OR overtime and PACU overtime. A samplegradient-based algorithm is proposed to solve the sample average approximation
of our formulation.
■ TA02
2 - Leveraging Predictive Analytics for HPC Scheduling in
Dynamic Environments
Sarah Powers, Oak Ridge National Laboratory, One Bethel Valley
Road, Oak Ridge, TN, United States of America,
powersss@ornl.gov
Optimization Applications in Homeland Security
Cluster: Homeland Security
Invited Session
Chair: Daniel Faissol, Lawrence Livermore National Laboratory,
Livermore, CA, United States of America, faissol1@llnl.gov
1 - Modeling the Global Spread and Impact of Diseases at Various
Levels of Aggregation
Daniel Skorski, Operations Research Scientist, Pacific Northwest
National Laboratory, 301 Hills Street, Richland, WA, 99352,
United States of America, Daniel.Skorski@pnnl.gov,
Robert Brigantic, Brent Daniel, Matthew Oster
Improvements in heterogeneous HPC scheduling can be obtained by leveraging
predictive analytics of job submissions. Development of the necessary workflow
models requires historical data and is costly due to the potential high diversity of
job types and their evolving patterns over time. We propose a method which
learns these patterns dynamically, allowing for unknown jobs types and changing
arrival patterns. Prediction gains are thus automated and utilizable in dynamic
environments.
Diseases spread by various modes of transportation is a never-ending modeling
and analysis need. GlobalCURE provides a framework to study the interplay
258
TA06
3 - Driver Scheduling Optimization Method Proposal for the J.B. Hunt
Intermodal Division
Luisa Janer, Graduate Student, University of Arkansas, 759 S
Royal Oak Pkwy #201, Fayetteville, AR, 72701, United States of
America, mjanerru@uark.edu, Valeria A. Remon Perez,
Nicole Taborga Delius, Nakia Lynn Lee
4 - Within and Cross-channel Effects of Brand Advertising
on Word-of-Mouth
Linli Xu, Carlson School of Management, University of
Minnesota, 321 19th Ave S, Suite 3-150, Minneapolis, MN,
United States of America, linlixu@umn.edu, Mitchell Lovett,
Renana Peres
A scheduling tool based on optimization was developed in order to improve the
driver and truck scheduling process of the J.B. Hunt Intermodal Division. After
having developed six prototypes of an optimization model, the tool manages to
effectively lower the outsourcing percentage to ten percent and increases the
driver-truck ratio to 1.8.
The central theme of this paper is to examine the relationship between
advertising and WOM. We study the influence of advertising on word-of-mouth
within channel and across channels. Preliminary evidence suggests significant
relationships both within and cross-channels. For example, both TV and Internet
display advertising appear to be significantly related to offline word-of-mouth
with TV having a stronger direct effect than Internet, whereas Internet advertising
is stronger online than TV.
■ TA04
5 - Mobile Big Data Analytics
Xueming Luo, Temple Univ, 1801 Liacouras, Philadelphia, PA,
19076, United States of America, luoxm@temple.edu
Panel Discussion: Journal Publication Tips
Over 3.6 billion people worldwide are deeply engaged with smartphone devices.
This reach potential proffers unprecedented marketing opportunity. As marketers
can send ads to smartphone users anywhere they are, marketing discipline now
faces tremendous opportunities of coming up with new theory and industry
practices for manager and consumer insights. Xueming will present some recent
research findings from his Global center for big data in mobile analytics.
Sponsor: Junior Faculty Interest Group
Sponsored Session
Chair: Cameron MacKenzie, Assistant Professor, Iowa State University,
3004 Black Engineering, Ames, IA, 50011, United States of America,
camacken@iastate.edu
1 - Panel Discussion: Successful Journal Publication Tips
Moderator: Cameron MacKenzie, Assistant Professor, Iowa State
University, 3004 Black Engineering, Ames, IA, 50011, United
States of America, camacken@iastate.edu, Panelists: Chris Tang,
Martin Savelsbergh, Serguei Netessine, Stefanos Zenios,
Jay Simon
■ TA06
Systemic Risk
Sponsor: Financial Services
Sponsored Session
Panel discussion will include editors and associate editors from Management
Science, Operations Research, Decision Analysis, Manufacturing & Service
Operations Management, and Transportation Science.
Chair: Stathis Tompaidis, Professor, University of Texas at Austin, Office
of Financial Research, Austin, TX, 78712, United States of America,
Stathis.Tompaidis@mccombs.utexas.edu
1 - Gauging form PF: Data Tolerances in Regulatory Reporting on
Hedge Fund Risk Exposures
Phillip Monin, Researcher, Office of Financial Research, 717 14th
St. NW, Washington, DC, 20005, United States of America,
Phillip.Monin@treasury.gov, Mark Flood, Lina Bandyopadhyay
■ TA05
Social Media and Networks in Business
Cluster: Social Media Analytics
Invited Session
We examine the precision of Form PF as an instrument for measuring risk
exposures in the hedge fund industry. Using a novel simulation methodology, we
assess the measurement tolerances of Form PF by examining the distribution of
actual portfolio risk exposures that are consistent with a fixed presentation on
Form PF. We find that Form PF’s measurement tolerances are sufficiently large to
allow private funds with dissimilar actual risk profiles to report similar risks to
regulators.
Chair: Xiaojing Dong, Associate Professor, Santa Clara University, 500
El Camino Real, Lucas Hall, Marketing, Santa Clara, CA, 95053, United
States of America, xdong1@scu.edu
1 - Predicting Social Influence Based on Dynamic
Network Structures
Mandy Hu, Assistant Professor, The Chinese University of Hong
Kong, CUHK Business School, Marketing, Shatin, Hong Kong PRC, mandyhu@baf.cuhk.edu.hk
2 - Systemic Risk: The Dynamics under Central Clearing
Agostino Capponi, Columbia, Mudd 313, New York, NY, 10027,
United States of America, ac3827@columbia.edu
We develop a tractable model for asset value processes of financial institutions
trading with one central clearinghouse. Each institution allocates assets between
his loan book and his clearinghouse account. We show that a unique equilibrium
allocation profile arises when institutions adjust trading positions to hedge risks
stemming from their loan books. The stochastic dynamic equilibrium path shows
a buildup of systemic risk manifested through the increase of market
concentration.
This study examines how network structure and dynamics interplay with the
effect of social influence to facilitate diffusion. The context we consider is the
diffusion of a new smartphone from a major wireless carrier in two medium-sized
cities in China. We are able to identify the two most significant network measures
related to social influence are diversity of connection and time variation of edge
numbers. Our findings provide foundation on the network-based targeting
strategy.
2 - Matrix Metrics: Network-based Systemic Risk Scoring
Sanjiv Das, William And Janice Terry Professor Of Finance, Santa
Clara University, Leavey School of Business, 500 El Camino Real,
Santa Clara, CA, 95053, United States of America, srdas@scu.edu
3 - Hidden Illiquidity with Multiple Central Counterparties
Kai Yuan, Columbia Business School, 3022 Broadway, 4J,
Uris Hall, New York, United States of America,
kyuan17@mail.gsb.columbia.edu, Paul Glasserman,
Ciamac Moallemi
I develop a network-based systemic risk score that depends on individual risk at
each financial institution and interconnectedness across institutions. This risk
metric is decomposable into risk contributions from each entity, forming a basis
for taxing each entity appropriately. Spillover risk determines the scale of
externalities that one institution might impose on the system. Splitting up toobig-to-fail banks from the system does not lower systemic risk.
Convex margin requirements from CCPs create an incentive for a swaps dealer to
split its positions across multiple CCPs, effectively “hiding” potential liquidation
costs. To compensate, each CCP needs to set higher margin requirements than it
would in isolation. In the case of linear price impact, we show that a necessary
and sufficient condition for the existence of an equilibrium is that the two CCPs
agree on liquidity costs and a difference in views can lead to a race to the bottom.
3 - Motivation of User-Generated Content in a Social Network
Xiaojing Dong, Associate Professor, Santa Clara University, 500 El
Camino Real, Lucas Hall, Marketing, Santa Clara, CA, 95053,
United States of America, xdong1@scu.edu
This study focuses on understanding the motivation of user-generated content in
open-source environments and online social networks. In our data, to encourage
members to contribute more reviews on the site, the community introduced cash
payment to those who offered reviews. We the find the effect of such reward
actually depends on the level of social connectedness. Those with fewer
connections responded positively to the reward, and those with more connections
responded negatively.
259
TA07
■ TA07
2 - “If at First You Don’t Succeed”: Understanding Serial
Entrepreneurs on Kickstarter
Hallie Cho, INSEAD, 1 Ayer Rajah Avenue, Singapore, Singapore,
hallie.cho@insead.edu, David Clough
Pricing and Risk Modeling in Financial Engineering,
Operations Research
From the crowdfunding platform Kickstarter, we have data on 27,399 technology
and design projects created by 6960 entrepreneurs—1376 of whom are serial
entrepreneurs. We examine characteristics of the projects and the entrepreneurs
to understand what distinguishes a serial entrepreneur from a one timer. For 779
of the serial entrepreneurs, their first projects were failures. We investigate how
serial entrepreneurs respond to setbacks and how their resource gathering
strategy changes over time.
Cluster: Risk Management
Invited Session
Chair: Hongzhong Zhang, Assistant Professor, Columbia University,
1255 Amsterdam Ave, New York, NY, 10027, United States of America,
hz2244@columbia.edu
1 - Counterparty Risk in a Heterogenous Random Network Model
Stephan Sturm, Worchester Polytechnic Institute,
ssturm@wpi.edu, Eric Schaanning
3 - Wisdom or Madness? Comparing Crowds with Expert Evaluation
in Funding the Arts
Ethan Mollick, Assistant Professor, U. Penn, 2000 Steinberg HallDietrich Hall, 3620 Locust Walk, Philadelphia, PA, 19004, United
States of America, emollick@wharton.upenn.edu, Ramana Nanda
We discuss the consequences of the central clearing mandate for OTC derivatives.
Analysing the expected total and potential future counterparty exposure in a
heterogeneous random graph network allows us to analyse the consequences of
central clearing in a realistic model calibrated to actual market data.
Drawing on a panel of experts and data from the largest crowdfunding site, we
examine funding decisions for proposed theater projects. We find significant
agreement between the funding decisions of crowds and experts. Our findings
suggest that crowdfunding can play a role in complementing expert decisions by
allowing projects the option to receive multiple evaluations and thereby lowering
the incidence of false negatives.
2 - On Minimizing Drawdown Risks of Lifetime Investments
Bin Li, University of Waterloo, 200 University Avenue West, M3
Building, Waterloo, Canada, b226li@uwaterloo.ca,
David Landriault, Dongchen Li, Xinfu Chen
4 - Crowdsourcing Exploration
Yiangos Papanastasiou, Haas School of Business, UC Berkeley,
Berkeley, CA, 94720, United States of America,
yiangos@haas.berkeley.edu, Nicos Savva, Kostas Bimpikis
We study a lifetime investment problem to minimize the risk of occurrence of
significant drawdowns. We examine two financial market models and closed-form
optimal strategies are obtained. Our results show that it is optimal to minimize
the portfolio variance when the fund value is at its historic high-water mark.
When the fund value drops, the fund manager should increase the proportion
invested in the asset with a higher instantaneous rate of return.
In an online review platform, information on the quality of alternative service
providers is both generated and utilized by the consumer population.
Inefficiencies arise from the fact that information is generated as a byproduct of
self-interested consumer choices, rather than with the benefit of future
consumers in mind. Within a multi-armed bandit framework, we study how such
inefficiencies relate to alternative policies of information-disclosure to the
platform’s users.
3 - Beating the Omega Clock: An Optimal Stopping Problem with
Random Time-horizon
Hongzhong Zhang, Assistant Professor, Columbia University, 1255
Amsterdam Ave, New York, NY, 10027, United States of America,
hz2244@columbia.edu, Neofytos Rodosthenous
We study the optimal stopping of a perpetual call option in a random timehorizon under exponential spectrally negative Levy models. The time-horizon is
modeled as the so-called Omega default clock, which is the first time the
occupation time of the underlying process below a level exceeds an independent
exponential random variable. We show that the shape of the value function
varies qualitatively with model parameters. In particular, we show the possibility
of two disjoint continuation regions.
■ TA09
Using Big Data Analytics for Technology Intelligence:
Methods and Cases to Gather Intelligence on
Technological Innovations
4 - Impact of Bayesian Learning and Externalities on
Strategic Investment
Wenxin Xu, University of Illinois, Urbana IL 61801, United States of
America, wxu9@illinois.edu
Sponsor: Technology, Innovation Management & Entrepreneurship
Sponsored Session
Chair: Tugrul Daim, Professor, Portland State University, P.O. Box 751,
Portland, OR, 97201, United States of America, ji2td@pdx.edu
1 - Business Partner Recommendation Based on Machine Learning
of Customer-Supplier Relationships
Yuya Kajikawa, Tokyo Institute of Technology, 3-3-6 Shibaura,
Minato-ku, Tokyo, Japan, kajikawa@mot.titech.ac.jp,
Naoko Matsuda, Yi Zuo
We investigate the interplay between learning effects and externalities in the
problem of competitive investments with uncertain returns. We find a region of a
war of attrition between the two firms in which the interplay between
externalities and learning gives rise to counterintuitive effects on investment
strategies and payoffs.
Business partnership is vital not only for business development but also
information sharing and collaboration for innovation. In this work, we modeled
customer-supplier relationships among firms using statistical learning model by
support vector machine to support firms to find plausible business partners. The
result showed prediction accuracy over 80% in average, but a variance was found
between different sizes of firms. We discuss the mechanism determining the
relationships.
■ TA08
e-Business Models
Cluster: Business Model Innovation
Invited Session
2 - The Circle of Innovation
Fred Phillips, Distinguished Professor, Yuan Ze University,
R60401, Building 6, No.135, Yuan-Tung Rd, Taoyuan,
Taiwan - ROC, fred.phillips@stonybrook.edu
Chair: Simone Marinesi, Wharton, 562 Jon M. Huntsman Hall,
3730 Walnut St, Philadelphia, PA, 19104, United States of America,
marinesi@wharton.upenn.edu
1 - Online Grocery Retailing
Elena Belavina, University of Chicago Booth School of Business,
5807 S Woodlawn Ave, Chicago, United States of America,
elena.belavina@chicagobooth.edu
There is a high-level feedback between technological innovation and social
change. Innovation brings about new products and services, and new ways of
using them. These in turn lead to new ways to interact and organize. The new
structures generate new unfilled needs, which are opportunities for still more
innovation. This changes how we classify innovations, how we should analyze
statistics, and our views of technology assessment, market segmentation, and
product development for sustainability.
Grocery delivery is a market that many try to conquer. Appropriate pricing is key
for success. There is little consensus among different players (at times even within
one firm operating in different locations) on what is the best pricing scheme. For
example, Amazon Fresh in Seattle is using per order pricing while in San
Francisco - subscription fee. We provide recommendation for the preferred
pricing scheme based on various characteristics (delivery logistics, demand
variability etc.).
3 - Technology Assessment: Case of Robotics for Power Applications
Tugrul Daim, Professor, Portland State University, P.O. Box 751,
Portland, OR, 97201, United States of America, ji2td@pdx.edu,
Judith Estep
This paper presents an integration of data analytics methods and expert judgment
quantification to evaluate multiple robotics technologies for the power utilities.
260
4 - Evaluating Technology Adoption in Emerging Regions: Case of
Smart Phone in Saudi Arabia
Fahad Aldhaban, Portland State University, P.O. Box 751,
Portland, United States of America, aldhaban@gmail.com,
Tugrul Daim
TA12
2 - Container Relocation Problem with Partial Information
Virgile Galle, PhD Candidate, MIT, Cambridge, MA,
vgalle@mit.edu, Cynthia Barnhart, Setareh Borjian,
Patrick Jaillet, Vahideh Manshadi
We introduce two new versions of the container relocation problem. First we
suppose that container departure times are only partially known and propose an
efficient branching algorithm using sampling and pruning to solve this problem.
Moreover, the second variation assumes that none of the departure times are
known in advance. In that case, we provide lower bounds to support the intuition
that the “lowest-height” policy is optimal in both static and dynamic case
This paper reviews the adoption factors of smart phones in emerging regions.
Saudi Arabia is studied as a case study. This presentation will cover the qualitative
part of the work. This part helped filter factors and finalize the survey instrument
3 - Taxi Assignment: Offline and Data-driven Online Optimization
Sebastien Martin, PhD Candidate, MIT, Operations Research
Center, MIT, 77 Mass Ave, Bldg E40-130, Cambridge, MA,
United States of America, semartin@mit.edu, Dimitris Bertsimas,
Patrick Jaillet
■ TA10
Contextual Factors Affecting eBusiness Initiatives
Sponsor: E-Business
Sponsored Session
This research focuses on taxi routing and assignment to customers: we optimize
the actions and revenues of a taxi fleet. We use MILPs and randomized algorithms
to solve to optimality the full-information version of the problem where demand
is known beforehand. Then, we extend these methods to make data-driven
online decisions. We apply our methods on the Manhattan network using actual
2013 yellow cabs demand data.
Chair: Frank MacCrory, Massachusetts Institute of Technology, MIT
Initiative on the Digital Economy, 355 Main Street - NE25-768D,
Cambridge, MA, 02142, United States of America, maccrory@mit.edu
1 - Social Media Usage Implications for Project Success, Political
Preferences, and Leisure Activities
Joseph Vithayathil, Assistant Professor, Washington State
University, Carson College of Business, Pullman, WA, 99164,
United States of America, joseph.vithayathil@wsu.edu,
John Kalu Osiri, Majid Dadgar
4 - Online Packing in the Random Time Arrival Model
Le Nguyen Hoang, Postdoctoral Associate, MIT, 77 Massachusetts
Ave, Cambridge, MA, 02139, United States of America,
lenhoang@mit.edu, Dawsen Hwang, Patrick Jaillet
Much interest has recently been given to online packing under a uniformly
random permutation of request arrivals. We propose a more general and realistic
setting, where, instead, requests arrive at random times. In particular, we do not
assume the number of requests to be known ahead of time, and we allow for
heterogeneity in the probability distributions of the random arrival times. We
present different online algorithms and discuss their respective competitive ratios.
We use a survey to empirically analyze the effect of social media usage on
workplace project success, political preferences, and leisure activities such as
shopping and television viewing behavior. This work adds to the emerging
literature on the impact of social media. We find weak association of social media
usage with project success, political preferences and leisure activities. Results are
interpreted using social presence and media richness theories, and implications
are discussed.
■ TA12
2 - Content Pricing Strategies under Dual Medium Access
Ran Zhang, UC Irvine, CA, ranz2@uci.edu, Shivendu Shivendu
12-Franklin 2, Marriott
Pricing information goods on physical and digital medium is a challenging
question for content providers. We develop an analytical model where consumers
are heterogeneous in both valuation for content and preference for medium. We
show that while offering both bundle of mediums and digital medium is optimal
under some market conditions, offering digital medium only is optimal under
other conditions. The optimal price for digital medium can decrease with marginal
cost of physical medium.
Convexification-based Algorithms for Solving
Quadratic and Polynomial Programs
3 - Incentives for Selective Information Sharing
Aditya Saharia, Associate Professor, Gabelli School of Business Fordham University, 113 W. 60th Sreet, New York, NY, 10023,
United States of America, saharia@fordham.edu
Chair: Jitamitra Desai, Professor, Nanyang Technological University,
50 Nanyang Avenue, Singapore, Singapore, jdesai@ntu.edu.sg
1 - Minimum Triangle Inequalities and Algorithms for 0-1 QCQPs
Jitamitra Desai, Professor, Nanyang Technological University, 50
Nanyang Avenue, Singapore, Singapore, jdesai@ntu.edu.sg,
Xiaofei Qi, Rupaj Nayak
Sponsor: Optimization/Mixed Integer Nonlinear Optimization and
Global Optimization
Sponsored Session
An increased transparency in inter-organizational systems does not make
members of a value chain equally better off. Individual members may then try to
influence other members’ decisions by introducing strategic ambiguity by not
collect demand information or by selectively share information with only some
downstream members.
We present a new class of minimum triangle inequalities (MINTI) for 0-1 QCQPs.
We prove that these inequalities are superior to the traditionally used triangle
inequalities, and offer several variations of these new cutting planes. We also
present an improved branch-and-bound algorithm that incorporates certain
properties from the MINTI cuts, and prove the efficacy of these cuts via our
computational results.
■ TA11
2 - Non-negative Polynomial and Moment Conic Optimization
Mohammad Mehdi Ranjbar, Rutgers, 100 Rockafeller Rd,
Rutgers Business School, Piscataway, NJ, 08854,
United States of America, 59ranjbar@gmail.com, Farid Alizadeh
Online Optimization with Integer Applications
Sponsor: Optimization/Integer and Discrete Optimization
Sponsored Session
Non-negative polynomial cone and its dual, moment cone, are non-symmetric
cones and extremely bad scaled. Then common primal-dual method will not be a
good algorithm to be used. Recently Nesterov has proposed a new predictorcorrector path-following method. Skajaa-Ye have proposed a Homogeneous
interior point method using Nesterov’s predictor-corrector path-following method
for some non-symmetric conic problem. We will extend that to non-negative
polynomial and moment conic programming.
Chair: Virgile Galle, PhD Candidate, MIT, vgalle@mit.edu
1 - Real-time Revenue Management under Partially
Learnable Demand
Dawsen Hwang, PhD Candidate, MIT, 77 Massachusetts Avenue,
32-D678, Cambridge, MA, 02139, United States of America,
dawsen@mit.edu, Le Nguyen Hoang, Vahideh Manshadi,
Patrick Jaillet
3 - Robust Sensitivity Analysis of the Optimal Value of
Linear Programming
Guanglin Xu, PhD Student, University of Iowa, 321 Finkbine Ln
Apt. 11, Iowa City, IA, United States of America, guanglinxu@uiowa.edu, Samuel Burer
We study a real-time revenue management problem where stochastic information
about the future demand is unknown a priori and can only be partially learned.
We develop adaptive and non-adaptive booking-limit policies parameterized by
predictability of the demand. In the two extreme cases of fully learnable and fully
unpredictable demand, we recover the known performance guarantees. Our work
bridges the gap between classical adversarial and stochastic demand models, and
defines value of learning.
We study sensitivity analysis in linear programming problems where general
perturbations in the objective coefficients and right-hand sides are considered.
This generality leads to non-convex quadratic programs (QPs) that are difficult to
solve in general. We investigate copositive formulations and tight semi-definite
relaxations of these QPs and validate our approach on examples existing in the
literature, as well as our own examples.
261
TA13
4 - Higher Rank-Order Semidefinite Cutting Planes for
Nonconvex QCQPs
Xiaofei Qi, PhD Student, Nanyang Technological University,
50 Nanyang Ave, Singapore, Singapore, xqi001@e.ntu.edu.sg,
Jitamitra Desai, Rupaj Nayak
2 - Near Optimal Ambiguity Sets in Distributionally
Robust Optimzation
Vishal Gupta, Assistant Professor, USC Marshall School of
Business, 3670 Trousdale Parkway, Bridge Hall 401 G,
Los Angeles, CA, 90089-0809, United States of America,
guptavis@usc.edu
We introduce a polynomial-time scheme to generate higher rank-order
semidefinite cutting planes that serve to tighten convex relaxations of nonconvex
quadratically constrained quadratic programs (QCQPs) and significantly improve
lower bounds. Suitably defined row-and-column based operations are used to
speed up the process of generating these cuts, and computational comparisons
across different types of relaxations shows the efficacy of these new cutting plane
strategies.
We assess the strengths of data-driven ambiguity sets in distributionally robust
optimization (DRO) by bounding the relative size of a candidate set to a specific,
asymptotically optimal set. We find popular ambiguity sets are much larger than
this asymptotically optimal set, suggesting current DRO models are overly
conservative. We propose new “near-optimal” sets that are only a constant factor
larger than the optimal set and satisfy the usual robustness properties.
3 - A Time Based Choice Model
Tauhid Zaman, MIT Sloan School of Management, 50 Memorial
Drive, Cambridge, MA, 02139, United States of America,
zlisto@mit.edu
■ TA13
We present a choice model which incorporates the time it takes the user to make
a decision. Our model assumes that the further apart two items are in terms of
user preference, the faster a decision is made. We conduct a set of online polls and
find that this model captures actual human behavior. We also show that using
this time based choice model can learn user preferences with high accuracy than
standard choice models for a fixed sample size.
Optimizing Sharing Service/Economy Under
Uncertainty
Sponsor: Optimization/Optimization Under Uncertainty
Sponsored Session
Chair: Siqian Shen, Assistant Professor, University of Michigan,
1205 Beal Avenue, Ann Arbor, MI, 48105, United States of America,
siqian@umich.edu
1 - Optimal Location Design of Carsharing Fleet under Uncertain
One-way and Round-trip Demands
Zhihao Chen, czhihao@umich.edu, Siqian Shen
■ TA15
Patient Scheduling in Health Care
Sponsor: Optimization in Healthcare
Sponsored Session
We allocate vehicles in a homogeneous carshare fleet to contracted locations, to
maximize the expected revenue from random demand for one-way and round
trip rentals. We use a spatial-temporal network and optimize both risk-neutral
and CVaR-based risk-averse two-stage stochastic programs with high demand
satisfaction rates. The two-stage problems are solved via branch-and-cut with
mixed integer rounding and we give insights on carsharing location design from
data reported by Zipcar in Boston.
Chair: Joseph Milner, Associate Professor Of Operations Management,
Rotman School of Management, University of Toronto,
105 St.George Street, Toronto, ON, M5S3E6, Canada,
Joseph.Milner@Rotman.Utoronto.Ca
1 - Dynamic Patient Scheduling for Multi-Appointment
Health Care Programs
Adam Diamant, Assistant Professor Of Operations Management,
Schulich School of Business, York University,
111 Ian Macdonald Boulevard, Toronto, ON, M3J1P3, Canada,
adiamant@schulich.yorku.ca, Fayez Quereshy, Joseph Milner
2 - Online Resource Allocation with Limited Flexibility
Xuan Wang, New York University, 44 West 4th Street,
Suite 8-154, New York, NY, 10012, United States of America,
xwang3@stern.nyu.edu, Jiawei Zhang, Arash Asadpour
We consider a general class of online resource allocation problems with limited
flexibility, where a type j request can be fulfilled by resource j or resource j+1, and
we call this limited flexibility the long chain pattern. The long chain has been
studied in process flexibility and has been shown to be very effective in coping
with demand uncertainty under offline arrivals. We provide preliminary results
that show the effectiveness of the long chain when the arrivals are online.
We investigate the scheduling practices of a multidisciplinary, multistage,
outpatient health care program with no-shows. We formulate the problem as a
Markov Decision Process and use approximate dynamic programming to find
policies to schedule patients to appointments. We examine the quality of our
solutions via structural results and compare them to a simulation of the clinic.
Our results applied to the operation of a bariatric surgery program at a large
tertiary hospital in Toronto, Canada.
3 - On-demand Staffing: Incentive Wage Contracts with
Guaranteed Fill Rates
Zhichao Zheng, Singapore Management University, Lee Kong
China School of Business, 50 Stamford Road, Singapore, 178899,
Singapore, danielzheng@smu.edu.sg, Tao Lu, Yuanguang Zhong
2 - Flexible Hospital-wide Patient Scheduling
Daniel Gartner, Carnegie Mellon University, 5000 Forbes Avenue,
Pittsburgh, PA, 15213, United States of America,
dgartner@andrew.cmu.edu, Rema Padman
We study the on-demand economy and its impact on labor market efficiency. We
consider n employers with uncertain and time-varying demands, and a platform
operator providing on-demand staffing services. We propose a novel fill ratebased allocation policy enabling the on-demand workforce to be shared efficiently
among employers. We propose a form of incentive contracts based on fill rate
guarantees, and show that our contracts can induce the system-wise optimality in
decentralized systems.
We study a patient scheduling problem with admission decisions, clinical
pathways, day and overnight hospital resources, ward and surgical team
assignment flexibility, and overtime considerations. We model the problem using
Mixed-Integer Programming and embed it in a rolling horizon planning to take
into account uncertain recovery times of and remaining resource capacity for
patients. We analyze the impact of flexibility and uncertainty on several metrics.
3 - Coordinated Scheduling for a Multi-station Healthcare Network
Ester Dongyang Wang, PhD Candidate, University of Texas, IROM
Dept., Austin, TX, United States of America, wdy@utexas.edu,
Douglas Morrice, Kumar Muthuraman
■ TA14
Data-driven Optimization
As the population ages, our healthcare industry must face the challenge of
increasing demand for care under constrained budget and resources. Our research
focuses on one of the central factors to the success of healthcare
reform–outpatient appointment scheduling. We develop a mechanism that
coordinates appointment scheduling among multiple services in a healthcare
network to improve access of care and reduce patient no-show rate. Our
approach has the potential to yield a global optimal solution.
Sponsored Session
Chair: Gah-Yi Vahn, Assistant Professor, London Business School,
Sussex Place, Regent’s Park, London, NW1 4SA, United Kingdom,
gvahn@london.edu
1 - Data-driven Estimation of (s, S) Policy
Gah-Yi Vahn, Assistant Professor, London Business School,
gvahn@london.edu
4 - Appointment Scheduling and Walk-in Strategies with
Unpunctual Patients
Mohamad Soltani, University of Alberta, PhD Office, Business
Building,, University of Alberta, Edmonton, AB, T6G 2R3,
Canada, soltani@ualberta.ca, Michele Samorani
I derive a tractable algorithm for computing the optimal (s,S) policy when the
decision maker has access to historical demand data. I show that this scheme
yields asymptotically optimal (s, S) policy and derive analytical characterisations
of confidence intervals, which is useful for operational decision-making.
It is commonly believed that clinics that schedule appointments have lower
patients’ waiting time and providers’ overtime than clinics that only allow walkins. However, if we consider patient unpunctuality, walk-in-only clinics may
achieve a higher performance. In this research, we investigate the conditions
under which each strategy is preferable.
262
■ TA16
TA18
petroleum industry.
2 - Embedding Resilience on Logistic and Supply Chain Networks
Jose Santivanez, Associate Professor, Universidad del Turabo, P.O.
Box 3030, Gurabo, PR, 00778, Puerto Rico,
santivanezj@suagm.edu, Emanuel Melachrinoudis
Disjunctive Conic and Optimization Problems
Sponsor: Optimization/Linear and Conic Optimization
Sponsored Session
This paper develops models for improving resilience to disruptions on critical
infrastructures such as logistics and supply chain networks through locational,
coverage, and path selection decisions. Network resilience is measured by the
ratio of the delivered amount of service over the total requested service when a
propagating disruption occurs. Availability of service depends on the capability of
the network to establish connectivity between service facilities and customers.
Chair: Julio Goez, Postdoctoral Fellow, Ecole Polytechnique Montreal
and GERAD, 2900 Boulevard Edouard-Montpetit, Montréal, QC, H3T
1J4, Canada, jgoez1@gmail.com
1 - A Generalized Trust Region Subproblem with Hollows and
Non-Intersecting Linear Constraints
Boshi Yang, The University of Iowa, 14 MacLean Hall, Iowa City,
IA, 52242, United States of America, boshi-yang@uiowa.edu,
Samuel Burer, Kurt Anstreicher
3 - Improving Supply Chain Network Resiliency with Preferential
Growth Decision Making
Ashley Skeete, PhD Fellow, Western New England University,
1215 Wilbraham Road, Springfield, MA, 01119, United States of
America, ashley.skeete@wne.edu, Julie Drzymalski
We study an extended trust region subproblem (eTRS) in which a nonconvex
quadratic function is minimized over a structured nonconvex feasible region: the
unit ball with r hollows (or holes) and m linear cuts. Under some nonintersecting assumptions, when r = 0 or when r = 1 and m = 0, it is known that
the eTRS has a tight, polynomial-time solvable conic relaxation. We show that the
conic relaxation is also tight for general r and m precisely when some nonintersecting assumptions are satisfied.
Network resiliency is the ability to maintain operations and connectedness under
the loss of some structures or functions. This research develops decision making
techniques in the supply chain context to improve resiliency of existing supply
chain networks as they grow with time. Consideration is given to factors such as
network topology, production requirements, the presence of redundancies and
cost.
4 - Hub Location-allocation for Combined Fixed-wireless and
Wireline Broadband Access
Ramesh Bollapragada, Professor, College of Business, San
Francisco State University, 1600 Holloway Avenue, San Francisco,
CA, 94132, United States of America, rameshb@sfsu.edu,
Uday Rao, Min Li, Junying Wu
2 - On Disjunctive Conic Cuts: When They Exist, When They Cut?
Mohammad Shahabsafa, Lehigh University, 14 Duh Dr, Apt. 221,
Bethlehem, PA, 18015, United States of America,
mos313@lehigh.edu
The development of Disjunctive Conic Cuts (DCCs) for MISOCO problems has
recently gained significant interest in the optimization community. Identification
of cases when DCCs are not existing, or not useful, saves computational time. In
this study, we explore cases where either the DCC methodology does not derive a
DCC which is cutting off the feasible region, or a DCC does not exist. Among
others, we show that deriving DCCs directly for p-order cone optimization
problems seems to be impossible.
This paper studies a telecommunications hub location model that includes the
classical capacitated facility location problem on a wireline network, as well as a
wireless network with technological as well as capacity constraints. There are
multiple wireline and wireless hub types, differing in costs and capacities. We
present a mathematical model to maximize network profit, build and test a quick
greedy heuristic with the optimal, and conduct sensitivity analysis using
representative data.
3 - Disjunctive Conic and Cylindrical Cut Management Strategies for
Portfolio Optimization Problems
Sertalp Cay, Lehigh University, 200 W Packer Ave, Bethlehem,
PA, 18015, United States of America, sec312@lehigh.edu,
Tamás Terlaky, Julio Goez
■ TA18
Disjunctive conic and cylindrical cuts lead significant positive impact while solving
Mixed Integer Second Order Cone Optimization (MISOCO) problems. The
decision for adding and removing these cuts should take depth of the cut and
structure of the problem into consideration. In this study, we explore strategies to
apply these novel cuts to discrete portfolio optimization problems within a
Branch-and-Conic-Cut software package. Preliminary results are provided to
compare these strategies.
Scientometric Data Analytics
Cluster: Modeling and Methodologies in Big Data
Invited Session
Chair: Dohyun Kim, Myongji University, Yongin, Korea,
Republic of, norman.kim@gmail.com
1 - Ranking Outliers in Patent Citation Network using Attributes and
Graph Structure
Ali Tosyali, Rutgers, the State University of New Jersey, Dept. of
ISE 96 Frelinghuysen Road, CoRE Building, Room 201,
Piscataway, NJ, United States of America,
alitosyali4778@gmail.com, Byunghoon Kim, Jeongsub Choi,
Byoung-yul Coh, Jae-min Lee, Myong K (MK) Jeong,
Andrew Rodriguez
4 - Novel Family of Cuts for SDP Relaxations for Some Classes of
Combinatorial Problems
Elspeth Adams, elspeth.adams@polymtl.ca, Miguel Anjos
k-projection polytope constraints (kPPCs) are a family of constraints that tighten
SDP relaxations using the inner description of small polytopes, as opposed to the
typical facet description. We examine the properties of kPPCs, methods for
separating violated kPPCs and their impact on the bounds in a cutting plane
framework. Problems satisfying the required projection property, such as the
max-cut and stable set problems, will be considered and results will focus on large
instances.
Being able to rank patents in outlierness is a crucial task for patent analysis. In the
past, existing general outlier ranking methods have been applied to patent data.
In this work, we propose a new outlier ranking method developed especially for
patents in attributed patent citation network. We utilized both graph structure
and attributes to rank outlier patents in patent citation network.
■ TA17
2 - Scientometric Analysis of Carbon Capture and Storage Research
Faezeh Karimi, Dr, University of Sydney, Project Management,
Sydney, 2006, Australia, faezeh.karimi@sydney.edu.au,
Rajab Khalilpour
Network Resilience and Applications
Sponsor: Optimization/Network Optimization
Sponsored Session
This study investigates the evolutionary trends of the international collaborations
among the research community of carbon capture and storage (CCS) by looking
at the collaboration network of countries publishing on CCS. The study elaborates
how both international collaboration network and knowledge structure of the
field have notably developed and interlinked over the years especially after 2005
during which almost 94% of the publications appeared.
Chair: Konstantin Pavlikov, University of Florida, 1350 N. Poquito
Road, Shalimar, FL, 32579, United States of America,
kpavlikov@ufl.edu
1 - Resilient and Structurally Controllable Supply Networks
under Disruptions
Amirhossein Khosrojerdi, The University of Oklahoma, 202 West
Boyd Street, Suite 218, Norman, Ok, 73071, United States of
America, akhosrojerdi@ou.edu, Farrokh Mistree, Janet K. Allen,
Krishnaiyan Thulsiraman
3 - Keyword Hierarchy Detection using Keyword Network Analysis
Dohyun Kim, Myongji University, Yongin, Korea, Republic of,
norman.kim@gmail.com, We Shim, Oh-jin Kwon,
June Young Lee, Sejung Ahn
A resilient supply network is one that has the ability to recover quickly from
disruptions and ensure customers are minimally affected. Designing the structure
of supply networks to be controllable is a way toward resilience. A three-stage
method is proposed to design a resilient and controllable supply network under
structural disruptions. The method is exercised using an example from the
We developed a keyword hierarchy detection algorithm using the keywork
network. Using the detection method, the hierarchy of keywords collected from
the same semantic field may be built.The keyword hierarchy detection method
can be used for a automatic preprocessing step to refine keywords in various topic
modeling methods.
263
TA19
4 - Integrated Variable Importance Assessment in Multi-stage
Manufacturing Processes
Gianluca Gazzola, Rutgers Center for Operations Research,
100 Rockafeller Road, Piscataway, NJ, 08854, United States of
America, ggazzola@scarletmail.rutgers.edu, Jeongsub Choi,
Myong K (MK) Jeong, Byunghoon Kim
information, or computational limitations. We recover missing interactions from
vast corporate networks using novel biclique clustering techniques to detect the
most significant edges, and hence provide new insights into power structures.
2 - Running Your Optimization Model on the Cloud with the IBM
CPLEX Studio IDE
Frederic Delhoume, Software Engineer, IBM, 9 Rue de Verdun,
Gentilly, 94253, France, delhoume@fr.ibm.com
We introduce a method for the assessment of variable importance in
manufacturing processes characterized by a hierarchy of technical relationships
between stage variables. Regression models of direct technical relationships and a
novel permutation measure are employed to quantify the local contribution of
every variable. Global contributions are finally obtained by integrating these local
assessments, based on the overall structure of indirect and direct technical
relationships in the process.
We will show how to easily run optimization models from the IBM CPLEX Studio
IDE. We will also demonstrate how to monitor the cloud service and get local
results from the remote optimization service. A REST API way of running models
on the cloud will be shown.
■ TA21
■ TA19
Medical Decision Making in Cancer Care
Computational Integer Optimization
Sponsor: Health Applications
Sponsored Session
Sponsor: Computing Society
Sponsored Session
Chair: Christine Barnett, University of Michigan, 1205 Beal Ave.,
Ann Arbor, MI, United States of America, clbarnet@umich.edu
1 - Predictive Modeling for Optimal Design of Cancer
Detection Protocols
Selin Merdan, University of Michigan, 1205 Beal Avenue,
Ann Arbor, MI, 48109, United States of America,
smerdan@umich.edu, Brian Denton
Chair: Yan Xu, Director, SAS, 100 SAS Campus Dr., Cary, NC,
United States of America, yan.xu@sas.com
1 - Recent Advances in the FICO Xpress MIP Solver
Michael Perregaard, Xpress Team, FICO, International Square,
Starley Way, Birmingham, B37 7GN, United Kingdom,
MichaelPerregaard@fico.com
Diagnosis of chronic diseases often involves expensive and invasive tests and
procedures. Predictive models can play an important role in determining the
optimal diagnostic protocol based on individual patient risk factors. We discuss an
approach for developing predictive models using clinical observational data that
suffers from common sources of bias such as low disease prevalence and missing
data. We illustrate the use of these models for optimization of prostate cancer
diagnostic protocols.
We will present some of the recent MIP advances in the FICO Xpress solver, with
an emphasis on how it is able to exploit the ever increasing core counts of
modern CPUs.
2 - The SAS MILP Solver: Current Status and Future Developments
Philipp Christophel, SAS Institute Inc., 100 SAS Campus Dr.,
Cary, NC, 27607, United States of America,
Philipp.Christophel@sas.com, Menal Guzelsoy, Imre Polik,
Amar Narisetty
2 - Model-based Calibration for Natural History Modeling
Jing Voon Chen, University of Southern California, Epstein Dept
of Indus & Sys Eng, Los Angeles, CA, United States of America,
jingvooc@usc.edu, Julie Higle
We give an overview of the current status of the SAS mixed integer linear
programming (MILP) solver that is part of the SAS/OR product. The focus will be
on describing recent implementation efforts for the MILP presolver as well as
future development directions.
A natural history (NH) model often requires calibration of unobservable model
parameters to fit observed data. Uncertainty in the data and in the calibrated
parameters impacts confidence in the optimal decision. We propose a method for
model-based calibration that is resilient to these uncertainties, especially for
comparative analyses of disease screening or treatment strategies. Illustrative
examples and sensitivity analyses will be discussed.
3 - Performance Improvements and New Features in the
Gurobi Optimizer
Chris Maes, Senior Developer, Gurobi Optimization, Inc., 125
Beacon St, Apt. #4, Boston, MA, 02116, United States of America,
maes@gurobi.com
3 - Assessment of Individualized Human Papillomavirus (HPV)
Vaccination Strategies
Fan Wang, University of Arkansas, 4207 Bell Engineering Center,
Fayetteville, AR, United States of America,
fxw005@email.uark.edu, Shengfan Zhang
This talk will cover the latest developments in the Gurobi Optimizer. We’ll discuss
the new Gurobi Cloud, which makes it easy to launch one or more Gurobi
machines when you need them. We’ll also talk about our upcoming release,
which includes significant performance enhancements and several new features.
The human papillomavirus (HPV) is the most common sexually transmitted virus
in the U.S. To prevent multiple cancers attributable to the HPV, HPV vaccine is
recommended for preteens and teens who have not been exposed to HPV. We
develop a simulation model for the optimal design of personalized HPV
vaccination program, which incorporates multiple social-behavioral and
demographic risk factors. The efficacy of the HPV vaccination program is
evaluated in terms of the HPV-related health outcomes.
4 - CPLEX Keeps Getting Better
Andrea Tramontani, CPLEX Optimization, IBM Italy, Via Martin
Luther King 38/2, Bologna, Italy, andrea.tramontani@it.ibm.com
We present some of the new features and algorithmic techniques that have been
recently added to IBM ILOG CPLEX Optimizer, and we give detailed benchmark
results that demonstrate the performance improvements achieved in latest CPLEX
versions.
4 - Tailoring CRC Screening Strategy for Different Age- and
Gender-specific Population Subgroups
Carolina Vivas, Purdue University, West Lafayette, IN 47906,
United States of America, cvivas@purdue.edu, Nan Kong, Robert
Klein, Thomas Imperiale
■ TA20
Standard guidelines for colorectal cancer (CRC) strategies do not consider
different age- and gender-specific subgroups for tailored screening
recommendations. Recent evidence suggests that men tend to face a higher risk of
developing advance adenomas earlier than women. We apply Design of
Experiments techniques to quantify the risk differences on CRC disease
progression. Model based cost-effectiveness analyses of various screening
strategies are conducted for different population subgroups.
Cloud Services and Applications
Cluster: Cloud Computing
Invited Session
Chair: Grace Lin, Data Analytic Technology and Applications (DATA),
Data Analytic Technology and Applications (DATA), Taipei, Taiwan ROC, gracelin@iii.org.tw
1 - Revealing Power Structures through Novel
Biclustering Approaches
Sabine Baumann, Professor Dr., Jade University, College of
Mgmt, Info., Tech., Friedrich-Paffrath-Str. 101, Wilhelmshaven,
22880, Germany, sabine.baumann@jade-hs.de, Oliver Eulenstein,
Christoph Wunck
Cloud and big data provide unprecedented access to massive interaction networks
of people and organizations. However, exploring such rich data environments
encounters equally extensive challenges: unreliable, incomplete or distorted
264
■ TA22
TA24
constructed many-servers asymptotic regime, when all the servers use the
Processor Sharing scheduling rule.
4 - Capacity of Information Processing Systems
Kuang Xu, Stanford University, United States of America,
kuangxu@stanford.edu, Laurent Massoulie
Two-Sided Matching Markets
Sponsor: Applied Probability
Sponsored Session
We study an information processing system where jobs are to be inspected by a
set of experts. Inspections produce noisy results depending on the jobs’ hidden
labels and the expert types, and an inspection occupies an expert for one time
unit. The manager’s objective is to assign inspections so as to uncover the jobs’
hidden labels, using a minimum number of experts. Our main result is an
asymptotically optimal inspection policy as the probability of error tends to zero.
Chair: Peng Shi, MIT Operations Research Center, 1 Amherst Street,
E40-149, Cambridge, MA, 02139, United States of America,
pengshi@mit.edu
Co-Chair: Yash Kanoria, Assistant Professor, Columbia University, New
York, United States of America, ykanoria@columbia.edu
■ TA24
Co-Chair: Itai Ashlagi, MIT, 100 Main St, Cambridge, MA, 02139,
United States of America, iashlagi@mit.edu
1 - On the Efficiency of Stable Matchings in Large Markets
Sangmok Lee, Univ of Pennsylvania, 3718 Locust Walk,
Philadelphia, PA, United States of America,
sangmok@sas.upenn.edu, Leeat Yariv
Intelligent Heuristics and Systems
Sponsor: Artificial Intelligence
Sponsored Session
We study the wedge between stability and efficiency in large one-to-one
matching markets. We show stable matchings are efficient asymptotically for a
large class of preferences. In these environments, stability remains an appealing
objective even on efficiency grounds, and monetary transfers are not necessary
for efficiency purposes. Nonetheless, for severely imbalanced markets, when
preferences entail sufficient idiosyncrasies, stable outcomes may be inefficient
even asymptotically.
Chair: Sam Thangiah, Professor/director, Slippery Rock University,
Artificial Intelligence and Robotics Lab, 250 ATS, Slippery Rock, PA,
16057, United States of America, sam.thangiah@sru.edu
1 - A New Mathematical Model for Pattern Recognition in the
Context of Feed Forward Neural Networks
Sam Findler, Slippery Rock University, 510 Campus Side Cir,
Slippery Rock, PA, 16057, United States of America,
srf5132@gmail.com
2 - Short Lists in Centralized Clearinghouses
Nick Arnosti, Stanford University, Stanford, CA,
United States of America, narnosti@stanford.edu
This paper sets out a new mathematical model for pattern recognition—in the
form of what I call Pattern Recognition Circuits (PRCs) and Pattern Recognition
Automata (PRAs). These new forms are given formal definitions and some
preliminary theorems are proven. Next, the model is applied to feed forward
neural networks, providing experimental grounding for the theory. Finally, the
place of this new mathematical model in the context of general computational
theory is discussed.
In the presence of frictions, participants in centralized clearinghouses generally
fail to list all acceptable match partners. As a consequence, mutually acceptable
pairs are left unmatched. The number of unmatched agents (and the happiness of
matched agents) depends crucially on the structure of correlations in participants’
preferences. This work identifies a fundamental tradeoff between match quality
and quantity, and uses this to offer guidance for the design of school choice
mechanisms.
2 - Automatic Construction of Relational Features with Dataconda
Michele Samorani, Assistant Professor, University of Alberta,
3-20F Business Building, University of Alberta, Edmonton, AB,
T6G 2R6, Canada, samorani@ualberta.ca
3 - How Much Choice is There in Two-sided Matching Markets?
Itai Ashlagi, MIT, 100 Main St, Cambridge, MA, 02139,
United States of America, iashlagi@mit.edu
We study the structure of two-sided random matching markets with tiers. Our
results provide insights on the amount of choice agents have in the core.
Traditional data mining and statistical techniques require a single table as input;
by contrast, I tackle the problem of findings patterns in a set of related tables
(Customers, Purchases, etc). This is made possible by Dataconda, a software freely
available to academics which automatically generates a large number of features
using information from all tables. In this talk, I will illustrate the benefits of this
approach through an example in retailing.
■ TA23
Sponsored Session
3 - Massively Parallel GPU Accelerated Genetic Algorithm for
Optimal Task Mapping in HPC Applications
Ramanan Sankaran, Computational Scientist, Oak Ridge National
Laboratory, P.O. Box 2008 MS 6008, Oak Ridge, TN, 37831,
United States of America, sankaranr@ornl.gov
Chair: Itai Gurvich, Professor, Kellogg School of Management,
Northwestern, 2001 Sheridan Rd., Evanston, IL, 60201,
United States of America, i-gurvich@kellogg.northwestern.edu
1 - Approximations to Non-stationary Diffusion Processes
Harsha Honnappa, Perdue University, West Lafayette, IN, United
States of America, honnappa@purdue.edu, Peter Glynn
Parallel applications on high performance computing (HPC) systems require
network and topology aware task placement to ensure performance and
scalability. We present a genetic algorithm for the quadratic assignment problem
(QAP) that utilizes thousands of GPU accelerated compute nodes allocated for the
application to compute an optimal task mapping in a few seconds. We show its
convergence characteristics and impact on real life physics simulations on Titan,
the most capable HPC system in the US.
Non-stationary diffusion processes emerge as limits to time inhomogeneous
queueing processes in appropriately defined ‘high intensity’ regimes. In general,
however, the transition densities of non-stationary diffusion processes are not
known in closed form. Thus, in this talk, we present analytical approximations to
expectations of these diffusion processes. This is joint work with Peter Glynn.
4 - A Nao Humanoid Robot System for Interacting with
Autistic Children
Sam Thangiah, Professor/director, Slippery Rock University,
Artificial Intelligence and Robotics Lab, 250 ATS, Slippery Rock,
PA, 16057, United States of America, sam.thangiah@sru.edu,
Daniel Martin, Michael Parnes, Mike Monfore, Stephen Fulton,
Zachary Kearney, Brian Atwell, Justin Cather, Andrew Rindt,
Sam Findler
Asymptotic Optimality in Processing Networks
2 - On the Control of Fork-join Networks
Erhun Ozkan, University of Southern California, Marshall School
of Business, Los Angeles, CA, 90089, United States of America,
eozkan@usc.edu, Amy Ward
The NAO humanoid robot is a two feet tall robot with two hands and two feet
with 25 degrees of freedom, cameras, microphones, sonar, infra-red and tactile
and pressure sensors. It has various communication devices and an Intel ATOM
processor. We describe a system implemented using the NAO robot to interact
with autistic children. The system is designed to interact with autistic children in
skill levels ranging from social to communication.
We study a prototypical fork-join network with two job classes and a shared
server that processes both job types. We show that a cmu-type static priority
policy is asymptotically optimal when the shared server is in some sense slow at
processing the more expensive jobs. Otherwise, a state-dependent slow departure
pacing control, under which the shared server sometimes gives priority to the less
expensive jobs, is asymptotically optimal.
3 - Insensitivity and Optimality of Load Balancing with Processor
Sharing Servers
Varun Gupta, Varun.Gupta@chicagobooth.edu, Neil Walton
We present some recent results and ongoing work on near-optimality and
insensitivity properties of shortest queue load balancing under a carefully
265
TA25
■ TA25
2 - Dynamic Supply Risk Management with Multisourcing,
Discretionary Selling, and Signal-based Forecast
Ting Luo, University of Texas at Dallas, 800 W Campbell Rd,
Richardson, TX, 75080, United States of America,
ting.luo@utdallas.edu, Long Gao, Nan Yang, Renyu Zhang
Online Information Intermediaries
Sponsor: Information Systems
Sponsored Session
We study a firm’s procurement and selling decisions in a multiclass demand and
multisupplier inventory system. The optimal procurement is driven by
multisourcing and intertemporal substitution and optimal selling is driven by
customer segmentation and intertemporal rationing; they are synchronized with
dynamic forecast for adaptive and resilient risk mitigation. We examine the
critical role of advance supply signals and understand when and how to use them.
Chair: Animesh Animesh, Associate Professor, McGill University,
1001 Rue Sherbrooke Ouest, Montreal, QC, h3a1g5, Canada,
animesh.animesh@mcgill.ca
1 - First-mover Advantage in Online Review Platform
Qianran Jin, McGill University, 1001 Sherbrooke Street West,
Montreal, Canada, qianran.jin@mail.mcgill.ca, Animesh
Animesh, Alain Pinsonneault
3 - Optimal Procurement Design for a National Brand Supplier in the
Presence of Store Brand
Xinyan Cao, PhD Student, University of Wisconsin - Milwaukee,
3202 N Maryland Avenue, Milwaukee, WI, 53202,
United States of America, xinyan@uwm.edu, Xiang Fang
While first-mover advantage has been widely studied at firm-level, our research
focuses on individual-level first-mover advantage in online review platform. We
study whether early reviews receive higher proportion of helpful votes than later
reviews. Our preliminary results show that early reviews are perceived to be more
helpful than later reviews. The first-mover advantage is greater for high
frequency reviewer than low frequency reviewer.
We consider a supply chain consisting of a national brand supplier and a retailer
which intends to develop its own store brand. We develop a game-theoretic
framework to analyze the strategic interaction between the two players in the
presence of asymmetric information.
4 - Duopolistic Procurement Contracts with Horizontal Information
Asymmetry
Hongyan Xu, Professor, Chongqing University,
School of Econ. and Bus. Administration, Chongqing, China,
xuhongyan@cqu.edu.cn, Yu Tang, He Huang
2 - What Makes Geeks Tick? A Study of Stack Overflow Careers
Lei Xu, McGill University, 855 Sherbrooke Street West, Montreal,
Canada, lei.xu2@mail.mcgill.ca, Tingting Nian, Luis Cabral
The success of a platform depends crucially on a thorough understanding of
motivations behind user participation. The identification has always been a
challenging task. We use a revealed preference approach to show that career
concerns play an important role in user contributions to Stack Overflow, the
largest online programming community. We show that career concerns explain
16% drop in answers activity after a job change. Robustness tests are conducted
to tease out alternative explanations.
We formulate a Cournot competition model of two chains where suppliers possess
private information of reliability and manufacturers may or may not share cost
information with the opponent. This paper under various scenarios aims to
examine the contract design and the interplay of horizontal information
asymmetry and vertical information asymmetry.
3 - The Dynamics of Online Referral Channels and E-commerce
Website Performance
Wenjing Duan, Associate Professor, The George Washington
University, 2201 G Street, NW, Washington, DC, 20052,
United States of America, wduan@gwu.edu, Jie Zhang
■ TA27
Application-motivated Theories and Methods for
Multiobjective Optimization
This study investigates the dynamic relationship between three referral channels
—- search engine, social medial, and third-party advertising —- and e-commerce
website performance. Our results derived from vector autoregressive models
suggest a significantly differential predictive relationship between referrals from
the three channels and sales performance measures.
Sponsor: Multiple Criteria Decision Making
Sponsored Session
4 - The Interactions Between Herding and Social Media Word-ofMouth: Evidence from Groupon
Xitong Li, Dr., HEC Paris, 1 Rue de la Liberation, Batiment V,
2eme etage, Bureau 207, Jouy-en-Josas,, 78351, France,
lix@hec.fr, Lynn Wu
Chair: Margaret Wiecek, Department of Mathematical Sciences,
Clemson University, Clemson, SC, 29634, United States of America,
wmalgor@clemson.edu
1 - Preference Preservation in Inverse Multi-objective
Convex Optimization
Taewoo Lee, University of Toronto, 5 King’s College Road,
Toronto, Canada, taewoo.lee@mail.utoronto.ca, Timothy Chan
This study aims to test if there is any complementary interaction between herding
and social media WOM. Using a panel data set from Groupon.com, we show they
reinforce each other in driving product sales. To explore the underlying
mechanisms behind the complementarities, we find the herding effect is more
salient for experience goods than for search goods, but the effect of Facebookmediated WOM does not significantly differ between the two product categories.
We present a new inverse optimization model for convex multi-objective
optimization that accommodates any input solution and determines a nonzero
weight vector that preserves the original preference of the decision maker who
generated the solution. We demonstrate how a linear approximation to the model
and a successive linear programming algorithm can trade-off between preference
preservation and computational efficiency, using data from prostate cancer
radiation therapy.
■ TA26
2 - Biobjective Robust Optimization Problem over the Efficient Set to
Aid Decision Making
Daniel Jornada, Texas A&M University, 1700 Research Pkwy,
280B Schlumberger Bldg, College Station, TX, 77843,
United States of America, djornada@tamu.edu, Jorge Leon
Optimal Sourcing, Procurement Design, and
Eco-label System in Supply Chain Management
Cluster: Operations/Marketing Interface
Invited Session
We present a biobjective robust optimization formulation for identifying robust
solutions from a given Pareto set arising from a multiobjective program (MOP).
The objective functions consider both solution and model robustness when
decision values are subjected to uncertainty at the time of implementation. The
solution approach is based on facial decomposition. We illustrate the applicability
of the methodology to aid decision making in the area of energy planning.
Chair: Xiang Fang, Associate Professor, University of WisconsinMilwaukee, 3202 N Maryland Avenue, Milwaukee, WI, 53211, United
States of America, fangx@uwm.edu
Co-Chair: He Huang, Professor, Chongqing University,
School of Economics and Business Admin., Chongqing, China,
huanghe@cqu.edu.cn
1 - Eco-label System Impact on Market Share and Profit
Yu Xia, Associate Professor, Northeastern University, 214 Hayden
Hall, 360 Huntington Ave, Boston, MA, 02115, United States of
America, Y.Xia@neu.edu, Xu Yang, Shilei Yang
3 - Spatial Data for Multiobjective Shortest Path Analyses:
Small Decisions with Large Consequences
F. Antonio Medrano, Post Doctoral Researcher, University of
California at Santa Barbara, Santa Barbara, CA, 93106,
United States of America, medrano@geog.ucsb.edu,
Richard Church
Multiobjective shortest path analysis is often used for developing alternatives in
the engineering design of new infrastructure over terrain. While such analysis
may appear to be non-subjective, the decisions made in assigning costs from
features and in the connectivity of the raster network will have major impacts on
the number of solutions, their spatial configuration, and their objective values.
We discuss these factors and decisions when using GIS data, and their impacts on
the solution set.
This research works on the design of the eco-label and its impact on market share
and profit for the company that adopts the eco-label system. To design an ecolabel system, we need to determine number of levels of labels to structure and the
index standard of each level. The gaps between levels should be significant
enough to promote effort in producing greener product. In addition, reaching a
higher level will bring additional business benefit such as profit for the engaged
manufacturers.
266
TA30
4 - Distributed Computation of Pareto Sets
Margaret Wiecek, Department of Mathematical Sciences,
Clemson University, Clemson, SC, 29634, United States of
America, wmalgor@clemson.edu, Brian Dandurand
■ TA29
The needs of multidisciplinary engineering design have motivated the
development of a distributed solution approach to computing Pareto solutions to
nonconvex decomposable multiobjective optimization problems. Existing results
on augmented Lagrangian coordination techniques and the block coordinate
descent method are extended into the multiobjective setting. These convergence
analyses lead to a MultiObjective Decomposition Algorithm (MODA) that is
applied to packaging in automotive design.
Sponsor: Analytics
Sponsored Session
■ TA28
The term analytics emerged, and went viral, in November 2005. Since then, there
have been many different, and often contradictory, definitions proposed for
analytics. Questions such as whether or not analytics is a discipline, and what
type of relationship it has to disciplines such as statistics, computer science and
operations research, have been unanswered. In this talk, a framework will be
presented that explains the emergence of analytics, and logically relates it to other
disciplines.
Applications of Analytics I
Chair: Tarun Mohan Lal, Mayo Clinic, mohanlal.tarun@mayo.edu
1 - Analytics: A Conceptual Framework
Robert Rose, President, Optimal Decisions LLC, 4 Kirby Lane,
Franklin Park, NJ, 08823, United States of America,
optimaldecisions@mac.com
Auctions for Ad Space
Cluster: Auctions
Invited Session
2 - Spreadsheet Software for Linear Regression Analysis
Robert Nau, Professor, Duke University, Fuqua School of
Business, Durham, NC, 27708, United States of America,
robert.nau@duke.edu
Chair: Ian Kash, Microsoft Research, 21 Station Road, Cambridge,
United Kingdom, iankash@microsoft.com
1 - General Truthfulness Characterizations via Convex Analysis
Rafael Frongillo, Postdoctoral Fellow, Harvard University,
29 Oxford Street, Cambridge, MA, United States of America,
raf@cs.berkeley.edu, Ian Kash
Spreadsheet add-ins for statistical analysis vary widely in terms of their user
interfaces, the detail and design of their output, and support for best practices of
analysis. This talk will give a brief overview of some of the market-leading
products and compare their linear regression features with a free add-in, RegressIt
(http://regressit.com), which was originally designed for teaching an advanced
elective on forecasting at Duke University and is now publicly available and
widely used.
We present a model of truthful elicitation which generalizes and extends
mechanisms, scoring rules, and related settings. Our main result is a
generalization of previous characterizations, including a new one for scoring rules
on non-convex sets of distributions. We generalize this model to eliciting some
property of the agent’s private information, and provide the first general result for
this setting. We also show how this yields a new proof of a result in mechanism
design due to Saks and Yu.
3 - Strategic Research: Analytics Excellence
Charity Maynard, Senior Health System Engineer, Mayo Clinic,
200 First Street SW, Rochester, MN, 55905, United States of
America, Maynard.Charity@mayo.edu
2 - Inefficiency and Low Revenue in GSP Auctions
Indranil Chakraborty, indro@nus.edu.sg
Mayo Clinic has a long rich legacy of integrating analytics, operations research
and industrial engineering in clinical care, research and education for achieving
excellence. This presentation will highlight the legacy, sophisticated
infrastructure, novel application and dissemination of learning. Key success
factors for leveraging analytics and engineering to address the formidable
challenges in health care today and tomorrow will be discussed.
In symmetric-Nash equilibrium GSP auction is efficient and generates more than
VCG revenue. The revenue ranking may fail under Nash conditions but not when
the maximum GSP revenue is considered. We let bidders to have values from
conversions instead of clicks, and assume recency effect on conversion to show
that in a broad range of situations the GSP auction is inefficient. When efficient
equilibria exist the maximum efficient equilibrium revenue in many situations is
lower than the VCG revenue.
4 - Analytics to Support Innovation in Outpatient Care
Delivery Processes
Tarun Mohan Lal, Mayo Clinic, mohanlal.tarun@mayo.edu
3 - Optimising Trade-offs Among Stakeholders in Ad Auctions
David Kurokawa, PhD Student, Carnegie Mellon University,
Computer Science Department, 5000 Forbes Ave, Pittsburgh, PA,
15213, United States of America, dkurokaw@cs.cmu.edu,
Yoram Bachrach, Sofia Ceppi, Ian Kash, Peter Key
With the growing trend and concern surrounding health care workforce
shortages, there is an increasing call for the redesign of office practices to reduce
inefficiency and improve capacity through better use of existing office staff. In this
presentation, we will discuss some innovative models of care delivery such as
increase pre-visit work, non-face face visits being implemented at Mayo Clinic
that has potential for improved operational performance and staff satisfaction.
We examine trade-offs among stakeholders in ad auctions. Our metrics are the
revenue for the auctioneer, number of clicks for users, and welfare for advertisers.
We show how to optimize linear combinations of these utilities via a GSP auction
with a per-click reserve price. We then examine constrained optimization of these
utilities. Finally, we examine a richer setting that allow using the screen real
estate in various ways.
■ TA30
4 - Mechanism Design for Mixed Ads
Ian Kash, Microsoft Research, 21 Station Road, Cambridge,
United Kingdom, iankash@microsoft.com, Sofia Ceppi,
Reza Khani, Yoram Bachrach, Peter Key
The GSP auction works when ads are simple, but does not generalize to richer
settings. Truthful mechanisms, such as VCG do. However, a straight switch from
GSP to VCG incurs significant revenue loss. We introduce a transitional
mechanism which mitigates this revenue loss. The mechanism is equivalent to
GSP when nobody has updated their bid and is equivalent to VCG when
everybody has updated. Our mechanism is based on a technique for making
payment functions into a truthful mechanism.
Chair: Stephen Hill
Sports and Entertainment
Contributed Session
Assistant Professor, UNC Wilmington, 601 South College Road,
Wilmington, NC, 28403-5611, United States of America,
hills@uncw.edu
1 - Does the Number of Days between Professional Sports Games
Really Matter?
Keith Willoughby, University of Saskatchewan, 25 Campus Drive,
Saskatoon, SK, S7N 5A7, Canada, willoughby@edwards.usask.ca,
Trevor Hardy
In the sport of football, teams typically play at most one game during a week.
Leagues may insert “bye weeks” into the schedule for each team, thereby
permitting rest and recuperation for players. Teams may feel unfairly treated if
they have more (or less) rest than their opponents. We explored over ten years of
regular season results from the Canadian Football League to determine if the
number of days off experienced by a team between games impacts team
performance.
267
TA31
2 - Optimization of Resource use in Massively Multiplayer
Online Games
Betty Love, University of Nebraska at Omaha, UNO Mathematics
Dept., 60th & Dodge Sts., Omaha, NE, 68182, United States of
America, blove@unomaha.edu, Andrew Cockerill
2 - Data Mining Corrections Testing
John Guerard, Director of Quantitative Research,
McKinley Capital Management, LLC, 3301 C Street, Suite 500,
Anchorage, AK, 99503, United States of America,
jguerard@mckinleycapital.com, Harry Markowitz, Ganlin Xu
With over 400 million players worldwide, massively multiplayer online games
(MMOs) continue to be a popular source of online recreation. MMOs frequently
involve resource management and virtual economies. This project demonstrates
the introduction of optimization strategies in the MMO game World of Warcraft.
A simulated annealing algorithm was implemented in a Lua script which runs in
the game’s user interface and determines how to use the player’s current
resources to maximize virtual profit.
Data mining corrections (DMC) tests of Global, Russell 3000, Non-U.S. stocks,
Emerging Markets, Japan-only, and China-only during the 2000-2014 period for
21 individual financial variables and two composite (robust, PCA-based)
regression models. We find that earnings forecasting models and regression-based
models emphasizing forecasted earnings acceleration and price momentum
models dominate the DMC tests which allow us to statistically dismiss Data
Mining as a potential source of modeling bias.
3 - A Gravity Model for Tourist Forecasting at FIFA Soccer
World Cups
Ghaith Rabadi, Associate Professor, Old Dominion University,
2102 Eng Systems Build, Dep. of Eng.Mngt. and Systems Eng.,
Norfolk, VA, United States of America, grabadi@odu.edu,
Mohammed Al-salem, Ahmed Ghoniem
3 - Applications of Machine Learning over Alpha Signals to Improve
Stock Selection and Boost Returns
Abhishek Saxena, Quantitative Research Analyst, McKinley
Capital Management, LLC, Suite 500, 3301 C Street,
Anchorage, AK, 99503, United States of America,
asaxena@mckinleycapital.com, Sundaram Chettiappan
FIFA Soccer World Cups are sport mega-events that enjoy tremendous popularity
worldwide. This paper analyzes historical bilateral tourist flows over the last two
decades to forecast the number of inbound tourists into future World Cup host
countries. Hosting sport mega-events will be considered as one of the input
factors to measure their impact on the number of tourists forecasted.
The paper explores the possibility of enhancing an alpha model through various
machine learning techniques. We show that these techniques can have
statistically significant additions to both raw returns and simulated returns in
various equity universes. These excess returns are mostly attributed to improved
stock selection as the risk profile doesn’t change significantly in terms of both
direct risk measurements (standard deviation based risk models) and exposures to
various fundamental factors.
4 - Optimal Hiking: Bi-modal Variation of the Traveling
Salesperson Problem
Roger Grinde, Associate Professor, University of New Hampshire,
Paul College of Business & Economics, 10 Garrison Avenue,
Durham, NH, 03824, United States of America,
roger.grinde@unh.edu
4 - The Rise of the Machines: Machine Learning in Stock Selection
Rochester Cahan, rcahan@empirical-research.com
Models that attempt to forecast the cross-section of future stock returns are often
structured as linear multifactor models. In this research we study the efficacy of
non-linear modeling techniques in stock selection strategies. We use a range of
factors known to predict stock returns as raw ingredients and investigate whether
various non-linear and machine learning algorithms can combine those
ingredients into predictive alpha signals, using only information known ex ante.
We benchmark the predictive power of the non-linear models against traditional
linear regression models constructed using the same data and estimation
windows.
The problem addressed is motivated by a mountaneering problem where there is
a network a peaks (destinations) connected by trails and a network of parking
areas connected by roads. Various objectives are possible; generally one wishes to
construct a series of hikes that together visit all the destinations. A formulation
and solution approach is presented.
5 - Analysis of Potential Solutions to Competitive Imbalance
in the NBA
Stephen Hill, Assistant Professor, UNC Wilmington,
601 South College Road, Wilmington, NC, 28403-5611,
United States of America, hills@uncw.edu
■ TA32
The National Basketball Association (NBA) is in the midst of an extended period
of competitive imbalance with teams in the Western Conference widely viewed as
being stronger than those in the Eastern Conference. In this work, we evaluate a
set of possible changes to the structure of the NBA. Each of these changes is
analyzed via Monte Carlo simulation with the impacts on competitive balance
and playoff participation described.
Principles in Applied Probability
Sponsored Session
Chair: Josh Reed, Associate Professor, NYU, 44 W. 4th St., New York,
NY, 10012, United States of America, jreed@stern.nyu.edu
1 - Relating Busy Period Duration and the Single Big Jump Principle
in Heavy Traffic
Bart Kamphorst, PhD Student, CWI, Science Park 123,
Amsterdam, 1098 XG, Netherlands, B.Kamphorst@cwi.nl,
Bert Zwart
■ TA31
Financial Applications of Data Mining and Machine
Learning Techniques
Queueing literature shows many results for the M/G/1 queue with a fixed server
utilization. However, in practice the server utilization may be increasing due to a
growing number of jobs per time unit. This causes a significant increase in waiting
times and the busy period duration. I will present asymptotic relations for the tail
probabilities of the former characteristics. Moreover, I will illustrate a typical long
busy period and discuss its relation with the Principle of a Single Big Jump.
Sponsor: Data Mining
Sponsored Session
Chair: John Guerard, Director Of Quantitative Research, McKinley
Capital Management, LLC, 3301 C Street, Suite 500, Anchorage, AK,
99503, United States of America, jguerard@mckinleycapital.com
1 - Optimal Global Efficient Portfolio with Emerging Markets using
Earning Forecasts
Shijie Deng, Georgia Inst of Tech, 755 Ferst Dr, Atlanta, GA,
United States of America, sd111@gatech.edu
2 - Capacity Allocation in a Transient Queue
Britt Mathijsen, PhD Student, Eindhoven University of
Technology, P.O. Box 513, 5600 MB, Eindhoven, Netherlands,
b.w.j.mathijsen@tue.nl, Bert Zwart
We consider an optimal capacity allocation problem of a two-period queueing
model, being in steady-state in the first time interval, but changing parameters at
the instance of the new period. The error in the objective function made by
disregarding the transient phase before reaching stationarity in this second
interval is quantified and approximated. Furthermore, we analyze the
consequence of staffing the system according to its steady-state behavior and
propose a corrected staffing rule.
We apply a multi-factor stock selection model which includes earning forecast to
analyze the performance of the optimal global portfolio which includes the
emerging markets. Under the Markowitz mean-variance framework, applied
optimization techniques are employed to address the practical issues of risktolerance, turn-over, and tracking-error. The impacts of these practical constraints
on the portfolio performance are analyzed through extensive numerical
experiments.
3 - Analysis of Cascading Failures
Fiona Sloothaak, PhD Student, Eindhoven University of
Technology, P.O. Box 513, 5600 MB, Eindhoven, Netherlands,
f.sloothaak@tue.nl, Bert Zwart
Inspired by analyzing the reliability of energy networks, particularly the
occurrence of large blackouts, we consider a stylized model of cascading failures.
By using connections with extreme value theory and Brownian bridge
approximations, we establish that the number of failed nodes follow a power law.
Time permitting, we also discuss connections with similar models and questions
from material science.
268
4 - On a Class of Reflected AR(1) Processes
Josh Reed, Associate Professor, NYU, 44 W. 4th St., New York, NY,
10012, United States of America, jreed@stern.nyu.edu,
Michel Mandjes, Onno Boxma
■ TA34
We study the the recursion Z(n+1) = max(aZ(n) + X(n),0) when X(n) is i.i.d. with
distribution the same as the difference of a positive random variable and an
independent, exponential random variable. We find the transform of Z(n) and,
when |a|<1, we perform a stationary analysis. In heavy-traffic, we show that the
process converges to a reflected Ornstein-Uhlenbeck process and the steady-state
distribution converges to the distribution of a normal random variable
conditioned to be positive.
Sponsored Session
TA35
Operations in Emergency Medicine
Chair: Yu Wang, PhD Student, Indiana University, yw39@indiana.edu
Co-Chair: Alex Mills, Assistant Professor, Indiana University, 1309 E.
10th Street, Bloomington, IN, 47405, United States of America,
millsaf@indiana.edu
1 - Coordinated Response of Health Care Networks in Mass
Casualty Incidents
Mercedeh Tariverdi, PhD Student, University of Maryland,
mercedeh@umd.edu, Elise Miller-Hooks, Thomas Kirsch,
Scott Levin
■ TA33
Medical Decision Making in Chronic Disease
Screening and Treatment
A hybrid analytical-simulation and system-based approach is presented for
assessing the benefits of coordinated response of a health care network in a mass
casualty incident. The method accounts for incident-related operational
disruptions along with other sources of transient system behavior. Critical
resource management is included.
Sponsored Session
Chair: John Silberholz, PhD Student, MIT, 77 Mass Ave, Bldg E40-130,
Cambridge, MA, 02139, United States of America, josilber@mit.edu
1 - An Analytics Approach to Designing Combination Chemotherapy
Regimens for Cancer
Dimitris Bertsimas, Professor, MIT, 77 Massachusetts Ave.,
Cambridge, MA, 02139, United States of America,
dbertsim@mit.edu, Allison O’hair, Stephen Relyea,
John Silberholz
2 - An Empirical Study of Patient Discharge Decisions in
Emergency Departments
Eric Park, Postdoctoral Associate, The University of British
Columbia, 2053 Main Mall, Vancouver, BC, V6T1Z2, Canada,
eric.park@sauder.ubc.ca, Yichuan Ding, Mahesh Nagarajan
We analyze the physician’s patient discharge decision in EDs. We study how
inpatient wards play a role as additional resources to the ED in the discharge
process. We study over 530,000 patient discharges in five Canadian EDs.
We present a data-driven approach for designing new chemotherapy regimens for
advanced gastric and breast cancer. Our approach combines (i) construction of a
large-scale database of clinical trial results, (ii) statistical modeling to predict
outcomes of new drug combinations, and (iii) optimization models to select novel
treatments that strike a balance between maximizing patient outcomes
(exploitation) and learning new things about treatments that may be useful in the
future (exploration).
3 - Allocation Models for Cooperation between Ambulance Services
Lavanya Marla, Assistant Professor, University of Illinois at
Urbana-Champaign, 104 S. Mathews Avenue, 216E, Urbana, IL,
61801, United States of America, lavanyam@illinois.edu
We consider a setting where multiple ambulance service providers cooperate to
serve a population. Such settings have been observed in the case of large
casualties; and in emerging economies where 911-type services compete with
existing ad-hoc services. We first demonstrate the opportunity costs due to lack of
cooperation. Then we present a game-theoretic framework to model the
allocation of ambulances from competing service providers. We conclude with
results from a real-world case study.
2 - On Estimating Optimization Model Parameters in Health
and Medicine
Thomas Trikalinos, Associate Professor, Brown University,
thomas_trikalinos@brown.edu
Combining information from independent sources (meta-analysis) can increase
the likelihood of optimal actions in operational problems. Using as example the
optimization of breast cancer screening strategies, I will discuss methods for and
implications of synthesizing model parameter estimates from independent studies,
while accounting for biases (systematic errors) and nontransferability (differences
between the setting specified by the optimization and the settings of the data
sources).
4 - Surge: Smoothing Usage of Resources is Good for Emergencies
Yu Wang, PhD Student, Indiana University, yw39@indiana.edu,
Alex Mills, Jonathan Helm
Major hospitals often experience demand surges close to or above their capacity.
We study the interplay between reactive and proactive surge strategies and their
impacts on the hospital’s immediate response and recovery. We find that
immediate recourse actions at best sacrifice long-term recovery for short-term
capacity improvement, while proactive workload smoothing provides a Paretoimproving response in both short- and long-term operational performance.
3 - A Robust Approach to Designing Cancer Screening Strategies
John Silberholz, PhD Student, MIT, 77 Mass Ave, Bldg E40-130,
josilber@mit.edu, Dimitris Bertsimas, Thomas Trikalinos
Many models have been proposed to evaluate screening strategies for detecting
cancer. Though each model for some cancer could be used to identify effective
screening strategies, models’ assumptions and structures can vary dramatically,
leading to differing conclusions about the most effective strategy. Using robust
and stochastic optimization, we identify screening strategies that are effective
across multiple models, which could increase confidence in the quality of the
identified strategies.
■ TA35
Panel Discussion: Infusing Learning from Hospitality
and Service Design to Healthcare: A Panel Discussion
Cluster: Hospitality, Tourism, and Healthcare
Invited Session
4 - Prioritizing Hepatitis C Treatment in United States Prisons
Can Zhang, Georgia Institute of Technology, 499 Northside Cir
NW, Apt. 315, Atlanta, GA, 30309, United States of America,
czhang2012@gatech.edu, Anthony Bonifonte, Turgay Ayer,
Jagpreet Chhatwal, Anne Spaulding
Chair: Rohit Verma, Professor, Cornell University, School of Hotel
Administration, 338 Statler Hall, Ithaca, NY, 14853-6902, United States
of America, rohit.verma@cornell.edu
1 - Infusing Learning from Hospitality and Service Design to
Healthcare: A Panel Discussion
Moderator: Rohit Verma, Professor, Cornell University, School of
Hotel Administration, 338 Statler Hall, Ithaca, NY, 14853-6902,
United States of America, rohit.verma@cornell.edu, Panelists:
Craig Froehle, Nitin Joglekar
Correctional populations, which represent about 30% of the national Hepatitis C
virus (HCV) prevalence, offer a great opportunity to control the HCV epidemic.
New HCV treatments are very effective but also outrageously expensive.
Therefore, prisons are pressed to prioritize treatment decisions for HCV-infected
inmates. We propose a mathematical modeling framework for HCV treatment
prioritization decisions in prisons and present extensive numerical results based
on large datasets from US prisons.
While fundamentally different from each other, the Healthcare and Hospitality
industries also share many common characteristics, challenges and constraints.
The purpose of this session is to discuss if and how lessons learnt from hospitality
can be infused to design better services within the context of healthcare, wellness
and senior living.
269
TA36
■ TA36
We present an optimization model that combines hospitals’ nurse staffing
decisions with two classes of quick-response decisions: (i) adjustments to the
assignment of cross-trained nurses working the current shift in each unit and (ii)
transfers of patients between units and off-unit admissions. We use a simulation
to derive insights into the level of benefit that can be expected from integrating
the aforementioned quick-response methods in the staffing process.
Innovations on Disaster Response Logistics
Sponsor: Public Sector OR
Sponsored Session
3 - Analyzing the Relationship Between Two-phased Room Allocation
Policies in an Outpatient Clinic
Vahab Vahdatzad, PhD Candidate, Northeastern University,
360 Huntington Avenue, Boston, MA, 02215, United States of
America, vahdatzad.v@husky.neu.edu, James Stahl,
Jacqueline Griffin
Chair: Felipe Aros-Vera, arosvm2@rpi.edu
1 - Competition Over Funding Resources in Humanitarian Operations
Arian Aflaki, Doctoral Student, Duke University, 100 Fuqua
Drive, Box 90120, Durham, NC, 27708, United States of America,
arian.aflaki@duke.edu, Alfonso Pedraza-Martinez
This research analyzes the relationship between two phases of room assignment
in an outpatient clinic. Specifically, we studied the interplay between the use of
rooms for Medical Assistant and physicians during a patient visits. We
demonstrate that policies for assigning rooms to MA and physicians has a
significant impact on patient wait time and length of stay. Several room allocation
policies are examined using discrete event simulation and interactions between
two phases are investigated.
Donors seek control over their donations, while it hurts the operational efficiency
of Humanitarian Organizations (HOs). We model the trade-off between
operational performance, fundraising effort, and donor preferences and find that
HOs can benefit from limiting donors’ control over their donations; however,
competition forces HOs to give control to donors.
2 - Optimizing Humanitarian Logistics Concepts of Operations:
The Case of Haiti
Erica Gralla, Assistant Professor, George Washington University,
800 22nd Street NW, Washington, 20052, United States of
America, egralla@gwu.edu, Liam Cusack, Phillip Graeter
4 - An Agent-Based Simulation Model of HIV Transmission and
Control among Men who have Sex with Men in Baltimore City
Parastu Kasaie, Postdoctoral Fellow, Johns Hopkins University,
615 N. Wolfe st, E6039, Baltimore, MD, 21205,
United States of America, pkasaie@jhu.edu, David Dowdy
After a disaster, the Logistics Cluster coordinates logistics for various responding
humanitarian agencies. They must quickly set up a supply chain, determining
entry points, major transport corridors, storage hubs, and vehicle requirements.
This research supports these decisions by finding the minimum-cost supply chain
configuration. Results for the case of Haiti are presented.
We present an agent-based simulation model to project the population-level
impact of implementing HIV preventive therapy (PREP) and treatment (ART) for
high-risk men who have sex with men (MSM) in Baltimore city. We compare a
counterfactual scenario in which PrEP and ART continue to be used at current
(low) levels against scenarios in which different levels of coverage and adherence
are achieved. The primary outcome of interest is the HIV incidence among MSM
in Baltimore over five years.
3 - Assessment of Risk Management and Disaster Response
Capabilities through the Process Maturity Framework
Miguel Jaller, Assistant Professor, University of California, Davis,
One Shields Ave, Ghausi Hall, 3143, Davis, CA, 95616, United
States of America, mjaller@ucdavis.edu, Diego Suero,
Melissa Del Castillo, Nuris Calderón, Jose William Penagos
5 - Estimating the Energy Imbalance Characterizing the Rise in
Obesity Among Adults in England
Saeideh Fallah-Fini, California State and Polytechnic University,
Pomona, 3801 W. Temple Ave, Pomona, CA, 91768,
United States of America, sfallahfini@cpp.edu
This paper explores the Process Maturity Framework to assess the current state of
the risk management and disaster response capabilities of a region in a developing
country. Using data collected by the team in Colombia, the paper discusses the
results in terms of the maturity levels for the different factors that comprise the
processes of: risk management and understanding, risk mitigation, and disaster
management and response; and puts forward a number of achievable goals and
key practices.
This paper uses systems dynamics to present a population-level model that
quantifies the energy imbalance gap responsible for the obesity epidemic among
adults in England (across different gender and ethnicity subpopulations) during
the past two decades. The developed model also estimates the magnitude of
calorie reduction that should be targeted by obesity interventions to reverse the
current trajectory of the obesity epidemic.
4 - Adaptive Decision Making under Dynamic Information Update in
Limited Data Environments
Kezban Yagci Sokat, PhD Candidate, Northwestern University,
2145 Sheridan Road, C210, Evanston, IL, 60208, United States of
America, kezban.yagcisokat@u.northwestern.edu,
Irina Dolinskaya, Karen Smilowitz
■ TA39
39-Room 100, CC
Supply Chain Management and Marketing Interface
After a disaster, there is often limited information about infrastructure damage.
New data sources such as OpenStreetMap are emerging. Utilizing these new data
sources, we use clustering and various imputation techniques with pre-disaster
and post-disaster attributes to approximate incomplete information in a timely
manner for routing decisions.
Invited Session
Chair: Gangshu Cai, Santa Clara University, OMIS Department,
Lucas Hall 216N, Santa Clara, CA, 95053, United States of America,
gcai@scu.edu
1 - Effects of Demand Uncertainty and Production Lead Time on
Product Quality and Firm Profitability
Baojun Jiang, Olin Business School, Washington University in St.
Louis, MO, 63130, United States of America,
baojunjiang@wustl.edu, Lin Tian
■ TA37
Health Care Modeling and Optimization IX
Contributed Session
We study the effects of demand uncertainty and production lead time in a
distribution channel with one retailer outsourcing its production to one supplier.
We show that the supplier may have no incentive to improve its lead time even if
it is costless to do so. An increase in the supplier’s JIT production capacity can
lead to higher or lower product quality, benefiting the retailer but potentially
hurting the supplier. A better market can make both the supplier and the retailer
worse off.
Chair: Parastu Kasaie, Postdoctoral Fellow, Johns Hopkins University,
615 N. Wolfe St, E6039, Baltimore, MD, 21205,
United States of America, pkasaie@jhu.edu
1 - Dynamic Advance Overbooking with No-Shows and Cancellations
Van-Anh Truong, Columbia University, 500 West 120th St, New
York, NY, 10027, United States of America, vt2196@columbia.edu
We introduce the first tractable model of dynamic advance overbooking with noshows and cancellations. In this fundamental model, advance appointments must
be given to a stream of patients arriving randomly over time. Patients might
cancel or miss their appointments, with the likelihood of these events increasing
with their wait times.
2 - The Protection Economy: Problem Retention or
Problem Prevention?
Oded Koenigsber, Associate Professor, London Business School,
Regent’s Park, London, United Kingdom,
okoenigsberg@london.edu, Eitan Gerstner, Daniel Halbheer
2 - Integrating Quick-response Methods and Staffing Decisions
in a Hospital
Jan Schoenfelder, Research Assistant, Augsburg University,
Neusässer Strafle 47, Augsburg, Ba, 86156, Germany,
janschoe@indiana.edu, Daniel Wright, Edwin Coe,
Kurt Bretthauer
Companies are advised to invest in quality programs to solve and prevent
customer problems. This paper shows that profit-maximizing motivate companies
to peruse protection strategies under which customer problems are created or
preserved so that protection services can be offered to repair the damages created
through the problems. Thus, standard economic efficiency measures used in the
“solution economy” are inappropriate for the “protection economy”.
270
3 - Co-opetition in Services: The Boardwalk Phenomenon
Lucy Gongtao Chen, National University of Singapore, 15 Kent
Ridge Drive, Singapore, Singapore, bizcg@nus.edu.sg,
Tinglong Dai, Nagesh Gavirneni, Xuchuan Yuan
TA42
2 - Pareto Improving Flow Control Policies for Multi-server
Emergency Departments - New Perspectives
Hung Do, Assistant Professor, University of Vermont,
55 Colchester Ave., Kalkin Hall 207, Burlington, VT, 05405,
United States of America, hdo@bsad.uvm.edu, Masha Shunko
We consider two service firms (e.g. restaurants) that compete on price and
waiting time and cooperate on entertainment effort that reduces the waiting cost
of the patrons. We study monopoly and duopoly settings and in the latter, we
consider both individual and joint entertainment efforts. We show that by
cooperating on entertainment, the competing service firms are able to achieve
efficiency levels equivalent to that of monopoly settings.
Using Emergency Medical Services setting as motivation, we design and analyze
flow control policies for service systems with N multiple-server queues. We focus
on policies that improve performance of the system and benefit all involved
entities. We propose new perspectives on performance measures, novel methods
to comparatively analyze flow control policies and reveal managerial insights that
help design such Pareto improving policies in practice.
3 - Adopting Best Practices: Public Relative Performance Feedback
as a Tool for Standardizing Workflow
Hummy Song, Harvard University, Wyss House,
Soldiers Field Road, Boston, MA, 02163, United States of
America, hsong@hbs.edu, Karen Murrell, Anita Tucker,
David Vinson
■ TA40
40- Room 101, CC
Behavioral Operations III
Contributed Session
In complex service systems, standardizing workflows (not processes) may be an
effective way to improve operational performance. We explore how public
disclosure of relative performance feedback (RPF) on individual workers’
processing times can help standardize workflow and improve productivity. We
examine the effect of public RPF on worker productivity and the extent to which
this varies by whether standardized processes are in place. We also explore
potential mechanisms driving these effects.
Chair: Ling Li, Professor, Department Chair of IT, Old Dominion
University, 2064 Constant Hall, Norfolk, VA, 23529,
United States of America, lli@odu.edu
1 - Newsvendor Decision with Multiple Reference Points
Feng Li, Dr., South China University of Technology, Wushan
Road, Guangzhou, China, fenglee@scut.edu.cn, Ying Wei
This paper studies how bottom line and status quo as reference profits influence
the newsvendor behavior and the optimal order quantity. Employing tri-reference
point theory, psychological value of the profit is regarded as gain, loss, or failure
based on the two benchmarks. We find that the presence of bottom line decreases
the optimal order quantity. In additon, the optimal order quantity may decrease
with the wholesale price and increase with the retail price.
4 - The Disposition Decision: Handoffs and End-of-shift Effects in an
Emergency Department
Robert Batt, Asst. Professor, Wisconsin School of Business, UWMadison, 975 University Ave., Grainger Hall, 5279, Madison, WI,
53706, United States of America, rbatt@bus.wisc.edu, Diwas Kc,
Bradley Staats, Brian Patterson
2 - Prediction on Network Public Opinion in Online Communities of
Different Age Structures
Tianjiang Boning, Master, Huazhong University of Science and
Technology, 1037 Luoyu Road, Wuhan, HU, 430074, China,
t.j.mcgrady@hotmail.com
We look at the effect of emergency department patients hand-offs on operational
variables such as length of stay, revisit rate, physician productivity. We also
examine what factors impact the probability of a patient being handed-off at the
end of a shift versus being dispositioned by the current doctor.
We get the evolution law of different age structures in different internet public
opinion events through simulation and analysis, and analyze the effect of youth
group, middle-aged group and elderly group in each community on internet
public opinion respectively, and also find the special role that the elderly group
plays during the public opinion evolution processes. In the end, we propose some
effective suggestions for Government according to simulation results.
■ TA42
42-Room 102B, CC
Operational Decision Making in Healthcare
Sponsor: Manufacturing & Service Oper
Mgmt/Healthcare Operations
Sponsored Session
3 - Employees’ Cyber Security Behavior and Information
Security Policy
Ling Li, Professor, Department Chair of IT, Old Dominion
University, 2064 Constant Hall, Norfolk, VA, 23529,
United States of America, lli@odu.edu, Li Xu, Wu He
Chair: Vishal Ahuja, Southern Methodist University, P.O. Box 750333,
Dallas, TX, United States of America, vahuja@smu.edu
1 - Impact of Severity-adjusted Workload on Health Status of
Patients Discharged from an ICU
Song Hee Kim, Assistant Professor, Marshall School of Business,
University of Southern California, Los Angeles, CA, United States
of America, songheek@marshall.usc.edu, Edieal Pinker,
Elizabeth Bradley, Joan Rimar
This research focuses on cybersecurity by theoretically defining the conceptual
domains of employees’ online security behavior and beliefs. We examined the
relative importance of 10 factors that will be used for developing new training
methods and materials to improve employee’s awareness and skills to defend
against cybersecurity risks, and investigated the relationship between the
availability of cybersecurity policy and individual employee’s behavior and beliefs
toward cybersecurity issues.
We examine whether workload has a direct impact on the health status of
patients discharged from ICUs, using data from two ICUs and a new measure of
patient acuity called the Rothman Index (RI). The RI is updated hourly in the
ICU, enabling us to track the health status of patients. Also, leveraging the RI, we
measure ICU workload in a novel way that takes into account not only the census
but also patient acuity, and study this severity-adjusted workload’s impact on the
patient disposition.
■ TA41
41-Room 102A, CC
Studies in Healthcare Productivity
Sponsored Session
2 - Evidence of Strategic Behavior in Medicare Claims Reporting
Hamsa Bastani, Graduate Student, Stanford University,
United States of America, hsridhar@stanford.edu, Joel Goh,
Mohsen Bayati
Chair: Robert Batt, Asst. Professor, Wisconsin School of Business, UWMadison, 975 University Ave., Grainger Hall, 5279, Madison, WI,
53706, United States of America, rbatt@bus.wisc.edu
1 - Mining for Content: A Study of E-visits
Hessam Bavafa, Assistant Professor, Wisconsin School of Business,
Madison, WI, United States of America, hbavafa@bus.wisc.edu
Upcoding is the practice where medical providers alter claims data to receive
increased reimbursement. Previous studies on detecting upcoding have been
limited by unobserved confounders (e.g. provider quality and patient risk). We
present a novel approach using a double regression that exploits state-level
variations in adverse event regulation and instrumental variables to provide
evidence of upcoding at a national scale. We also make several policy
recommendations for reducing upcoding.
We study the micro-structure of e-visits, electronic communications between
patients and providers through patient portals. The main promise of e-visits as a
new channel for providing primary care services is to decrease the number of
office visits and improve patient health. We examine detailed information about
the patients, providers, and e-visit details (e.g., timings and text of e-visits) to
establish a better understanding of e-visits.
271
TA43
3 - Screening for Hepatocellular Carcinoma: A Restless Bandit Model
Elliot Lee, University of Michigan, 1205 Beal Ave, Ann Arbor, MI,
48109, United States of America, elliotdl@umich.edu,
Mariel Lavieri, Michael Volk
4 - Attribution in Online Advertising under Markov Browsing Models
Antoine Desir, Columbia University IEOR department, 500 West
120th Street, Mudd 315, New York, NY, 10027, United States of
America, ad2918@columbia.edu, Vineet Goyal, Garud Iyengar,
Omar Besbes
Currently, all patients at risk for hepatocellular carcinoma (HCC) are screened
every six months. Recent medical discoveries have found a correlation between a
biomarker measured at each screening, and his/her risk of developing HCC. We
model the problem of simultaneously learning while allocating a limited number
of screening resources across a population as a restless bandit model. We prove
several structural properties of this problem, and ultimately derive a
corresponding optimal policy.
Web viewers are exposed to multiple ads across different websites before they
potentially make a purchase (leading to a conversion). In turn, a key question
facing the online advertising industry is that of attribution. We analyze attribution
based on Shapley values. While intractable in general, we provide
computationally tractable approximations to Shapley values under a general
Markov chain customer browsing behavior model and compare this attribution to
heuristics commonly used in practice.
4 - Enhancing FDA’s Decision Making using Data Analytics
Vishal Ahuja, Southern Methodist University, P.O. Box 750333,
TX, United States of America, vahuja@smu.edu, John Birge
■ TA44
Existing FDA surveillance methods are based on voluntary reporting or metaanalysis primarily geared towards identifying new/unknown adverse events. We
propose a statistically robust and evidence-based empirical approach that focuses
on evaluating specific drug-related adverse outcomes to aid in the FDA decisionmaking. We demonstrate our approach using a controversial black box warning.
Based on a large dataset from the Department of Veterans Affairs, we find that the
warning was not warranted.
44-Room 103B, CC
Pricing Issues in Revenue Management
Sponsor: Revenue Management and Pricing
Sponsored Session
Chair: Rene Caldentey, NYU, 44 W 4th St, New York, NY, 10012,
United States of America, rcaldent@stern.nyu.edu
■ TA43
Co-Chair: Ying Liu, Stern School of Business, New York University,
44 West 4th Street, KMC 8-154, New York, NY, 10012,
United States of America, yliu2@stern.nyu.edu
1 - Incorporating Online Customer Ratings in Pricing Decisions
Marie-claude Cote, Manager, Data Science, JDA Software Innovation Labs, 4200 Saint-Laurent #407, Montréal, QC, H2W
2R2, Canada, Marie-Claude.Cote@jda.com, Philippe Tilly,
Nicolas Chapados
43-Room 103A, CC
Measurement and Optimization in Online Advertising
Sponsored Session
Chair: Omar Besbes, Professor, Columbia University, Graduate School
of Business, New York, NY, 10027, United States of America,
ob2105@columbia.edu
Research have demonstrated that online customer ratings have a huge impact on
the decision to choose a product. In hospitality, where the product is a hotel room
for a length of stay, customers consult an increasing number of reviews prior to
booking.We will describe an approach to automatically incorporate online user
rating impact in the pricing decisions of a hospitality revenue management
system.
Co-Chair: Vineet Goyal, Associate Professor, Industrial Engineering and
Operations Research, Columbia University, 500 West 120th Street, 304
Mudd, New York, NY, 10027, United States of America,
vgoyal@ieor.columbia.edu
2 - On the Equivalence of Quantity Pre-commitment and
Cournot Games
Amr Farahat, Washington University in St. Louis, One Brookings
Drive, St. Louis, MO, 63104, United States of America,
farahat@wustl.edu, Hongmin Li, Tim Huh
Co-Chair: Garud Iyengar, Columbia University, S. W. Mudd 314,
500W 120th Street, New York, NY, United States of America,
garud@ieor.columbia.edu
1 - Advertiser Revenue Versus Consumer Privacy in
Online Advertising
Vibhanshu Abhishek, Carnegie Mellon University, 5000 Forbes
Avenue, Pittsburgh, PA, 15213, United States of America,
vibs@andrew.cmu.edu, Arslan Aziz, Rahul Telang
We establish sufficient conditions under which Cournot outcomes solve quantityfollowed-by-pricing games. Kreps and Scheinkman (1983) established this
connection for homogeneous product duopolies and Friedman (1988) for certain
differentiated product oligopolies under restrictive assumptions. Our research
provides conditions for more general differentiated product settings, including
multinomial logit models.
Increasing concerns around consumer privacy have questioned the value of
targeted advertising. In this paper we quantify the value of privacy-intrusive
information in targeted advertising. Using individual level browsing/purchase
data we find that using more privacy-intrusive information increases the accuracy
of prediction of purchases, but at a decreasing rate. Targeted advertising is also
effective in increasing purchase probability. In our data, restricting cookies
reduces sales by 14%.
3 - Pricing Policies for Perishable Products with Demand Substitution
Ying Liu, Stern School of Business, New York University, 44 West
4th Street, KMC 8-154, New York, NY, 10012, United States of
America, yliu2@stern.nyu.edu, Rene Caldentey
We study a monopolist’s optimal dynamic pricing policy for a family of substitute
perishable products. Customers arrive to the market according to an exogenous
stochastic process, each with a budget constraint. Upon arrival, each customer
first makes a decision among subfamilies that are differentiated by quality, and
then selects among horizontally differentiated products within the subfamily. We
characterize the optimal pricing policy and study the asymptotic approximation.
2 - Learning and Optimizing Reserve Prices in Repeated Auctions
Hamid Nazerzadeh, University of Southern California, Bridge
Memorial Hall, 3670 Trousdale Parkway, Los Angeles, 90089,
United States of America, hamidnz@marshall.usc.edu,
Yash Kanoria, Renato Paes Leme, Afshin Rostamizadeh,
Umar Syed
4 - Optimal Time and Price of Dynamic Upgrade
Xiao Zhang, PhD Candidate, University of Texas at Dallas,
Richardson, TX, United States of America,
xxz085020@utdallas.edu, Ozalp Ozer
A large fraction of online advertisements are sold via repeated second price
auctions. The reserve price is the main tool for the auctioneer to boost revenues.
However, the question of how to effectively set these reserves remains essentially
open from both theatrical and practical perspectives. The main challenge here is
that using previous bids to learn reserves could lead to shading of bids and loss of
revenue. I’ll present incentive compatible near-optimal leaning algorithms in this
context.
Upgrade, a strategy used in travel industry to balance the supply-and-demand
mismatches among products of different quality levels, is usually offered either at
the booking time or the consumption time. We study a revenue management
problem of a firm which sells two products at fixed prices and offers upgrade
options anytime when necessary. The optimal policy specifies the time and price
of the upgrade option, and how many existing customers should be offered this
option.
3 - Advertising on a Map
Sergei Vassilvitskii, Google, 111 8th Avenue, New York,
United States of America, sergei@cs.stanford.edu
We study the mechanism design problem for advertising on a map. Unlike
traditional search advertising where there is a linear order on the slots, no such
structure exists in the case of a map. We begin with a model of the setting, noting
that the utility of an ad is discounted by the presence of competing businesses
nearby and its position in the set of ads ordered by distance from the user. We
present simple, approximately welfare maximizing allocation schemes with good
incentive properties.
272
■ TA45
TA47
2 - Spatial Competition and Preemptive Entry in the Discount
Retail Industry
Fanyin Zheng, Columbia Business School, 3022 Broadway, New
York, NY, United States of America, fanyin.zheng@gmail.com
45-Room 103C, CC
Social Learning and Revenue Management
I study the competitive store location decisions of discount retail chains in this
paper. I model firms’ entry decisions using a dynamic duopoly location game and
allow stores to compete over the shopping-dollars of close-by consumers. I use
various economic modeling technics to make the model tractable and infer
market divisions from data using a clustering algorithm. The empirical analysis
suggests that dynamic competitive considerations are important in chain stores’
location decisions.
Sponsored Session
Chair: Costis Maglaras, Columbia Business School, New York, NY,
10027, United States of America, c.maglaras@gsb.columbia.edu
Co-Chair: Alireza Tahbaz-Salehi, Columbia Business School,
3022 Broadway, Uris Hall 418, New York, NY, 10023,
United States of America, alirezat@columbia.edu
1 - Monopoly Pricing in the Presence of Social Learning
Davide Crapis, Columbia Business School, 3022 Broadway,
New York, NY, 10027, United States of America,
dcrapis16@gsb.columbia.edu, Bar Ifrach, Costis Maglaras,
Marco Scarsini
3 - Using Real-time Operational Data to Increase Labor
Productivity in Retail
Marcelo Olivares, Assistant Professor, Universidad de Chile,
Republica 701, Santiago, Chile, molivares@u.uchile.cl
We develop a methodology to re-assign sales employees across departments in a
large retail store in order to improve productivity. Our method seeks to maximize
the effectiveness of labor by allocating employees to departments that require
inmediate assistance and where this assistance has a larger impact of sales. The
method combines empirical methods to measure the impact of assistance and
store operational data collected through video analytics to reassign employees in
real-time.
A monopolist offers a product to a market of consumers with heterogeneous
preferences. Consumers are uninformed about product quality and learn from
reviews of others. First, we show that learning eventually occurs. Then, we
characterize the learning trajectory via a mean-field approximation that highlights
how the learning process depends on price and heterogeneity. Finally, we solve
the pricing problem and show that policies that account for social learning
increase revenues considerably.
4 - Consumer Search and the Structure of Personal Networks
Raghuram Iyengar, Associate Professor, The Wharton School,
University of Pennsylvania, 3730 Walnut Street Suite 700,
Philadelphia, PA, 19104, United States of America,
riyengar@wharton.upenn.edu
2 - Networks, Shocks, and Systemic Risk
Alireza Tahbaz-Salehi, Columbia Business School, 3022
Broadway, Uris Hall 418, New York, NY, 10023, United States of
America, alirezat@columbia.edu, Daron Acemoglu, Asu Ozdaglar
We study how consumers’ information search for and purchase of new products
are affected by structure of their personal network. To address threats to internal
validity common in network studies, we conduct a randomized experiment in
which we manipulate the similarity of preference among consumers and their
network contacts. We estimate consumers’ utility function and determine how
network antecedents moderate the weight on others’ information.
We develop a unified framework for the study of how network interactions can
function as a mechanism for propagation and amplification of microeconomic
shocks. The framework nests various classes of games over networks, models of
macroeconomic risk originating from microeconomic shocks, and models of
financial interactions.
3 - Market Entry under Competitive Learning
Kimon Drakopoulos, kimondr@mit.edu, Asu Ozdaglar,
Daron Acemoglu
■ TA47
We consider a market entry game with two players, an incumbent and an
entrant. The market can be of two types: (a)bad in which case the demand is fully
elastic at a price \bar{p} or(b) good in which case there is a positive arrival rate of
consumers who are willing to buy at higher prices. The entrant is learning the
type of the market by observing the flow of payoffs.We prove that the problem
has the structure of a war of attrition game and study its weak perfect Bayesian
equilibria.
Sustainability in Food Supply Chains
47-Room 104B, CC
Sponsor: Manufacturing & Service Oper Mgmt/Sustainable
Operations
Sponsored Session
Chair: Erkut Sonmez, Assistant Professor, Boston College,
140 Commonwealth Ave, Fulton Hall, Chestnut Hill, MA, 02446,
United States of America, erkut.sonmez@bc.edu
1 - Supply Chain Analysis of Contract Farming
A. Serdar Simsek, Cornell ORIE, 282 Rhodes Hall, Ithaca, NY,
14853, United States of America, as2899@cornell.edu, Awi
Federgruen, Upmanu Lall
4 - Social Learning with Differentiated Products
Arthur Campbell, Associate Professor, Yale University, School of
Management, 135 Prospect Street, P.O. Box 208200, New Haven,
CT, 06520-8200, United States of America,
Arthur.Campbell@yale.edu
This paper embeds social learning in a model of firms producing differentiated
products. We consider how the structure of social relationships between
consumers influence pricing and welfare. The model considers how a variety of
characteristics of the social network influence these outcomes. It also serves to
highlight the challenges one faces in using metrics such as consumer awareness
and the sensitivity of demand to prices as measures of informational efficiency in
markets.
Contract farming sustains the operations of vulnerable farmers while better
positioning the manufacturers to manage their supply risks. In this setting, a
manufacturer who owns several production plants -each with a random demand
for the crop- selects the set of farmers that minimizes her expected procurement
and distribution costs before the growing season. We present two solution
methods to this problem. We applied our model to a company contracting with
hundreds of small farmers in India.
2 - Processed Produce: Introduction, Pricing, and Profit Orientation
Omkar Palsule-desai, Faculty, Indian Institute of Management
Indore, Rau Pithampur Road, Indore, India,
omkardpd@iimahd.ernet.in, Muge Yayla-Kullu, Nagesh Gavirneni
■ TA46
46-Room 104A, CC
Empirical Research in Supply Chains and
Service Management
We examine product characteristics and market dynamics to identify conditions
under which it is optimal to introduce the processed produce, and it should be
managed by cooperatives instead of private firms. We develop a mathematical
model capturing (i) competition between non-profit and for-profit firms, (ii)
consumers’ valuation discount, and (iii) product perishability. We provide ample
evidences to policy makers promoting processed products offered by cooperatives.
Sponsor: Manufacturing & Service Oper Mgmt/Service Operations
Sponsored Session
Chair: Marcelo Olivares, Assistant Professor, Universidad de Chile,
Republica 701, Santiago, Chile, molivares@u.uchile.cl
1 - Estimating Customer Spillover Learning of Service Quality:
A Bayesian Approach
Andres Musalem, Universidad de Chile, Republica 701, Santiago,
Chile, amusalem@duke.edu, Yan Shang, Jeannette Song
3 - Converting Retail Food Waste Into By-product
Deishin Lee, Assistant Professor, Boston College, 140
Commonwealth Ave,, Fulton Hall 344, Chestnut Hill, MA, 02467,
United States of America, deishin.lee@bc.edu, Mustafa Tongarlak
By-product synergy (BPS) is a form of joint production that uses the waste stream
from one (primary) process as useful input into another (secondary) process. We
investigate how BPS can mitigate food waste in a retail grocer setting, and how it
interacts with other mechanisms for reducing waste (i.e., waste disposal fee and
tax credit for food donation). We derive the retailer’s optimal order policy under
BPS, showing how it affects the amount of waste.
We propose a Bayesian framework for estimating customer “spillover learning,”
— the process by which customers’ learn from previous experiences of similar but
not necessarily identical services. We apply our model to a data set containing
shipping and sales historical records provided by a world-leading third-party
logistics company.
273
TA48
4 - Improving Food Bank Gleaning Operations: An Application in
New York State
Erkut Sonmez, Assistant Professor, Boston College, 140
Commonwealth Ave, Fulton Hall, Chestnut Hill, 02446,
United States of America, erkut.sonmez@bc.edu, Miguel Gomez,
Deishin Lee, Xiaoli Fan
We consider a firm’s supply diversification problem under supply random yield
and price sensitive demand. We study two pricing schemes: responsive pricing
and ex ante pricing. We characterize the sourcing decisions under each pricing
scheme and compare them to study the strategic relation between diversification
and responsive pricing.
2 - Risk Pooling under Price and Demand Uncertainty
Refik Gullu, Professor, Bogazici University, Industrial Engineering
Dept., Bebek, Istanbul, 34342, Turkey, refik.gullu@boun.edu.tr,
Nesim K. Erkip
We develop a stochastic optimization model to help food banks to improve their
gleaning operations. Gleaning refers to collecting food from what is left in the
fields after harvest, and donating the goods to food banks or pantries that service
food insecure individuals.
We consider purchasing and distribution decisions for a commodity whose price is
random and correlated with its demand. A model, where the purchasing decisions
of locations are pooled is proposed. We obtain the optimal purchase quantity,
time and quantity of allocation, and quantify the benefits of pooling price risk.
■ TA48
3 - Bunching Supply Contracts with Information Asymmetry in a
Two-echelon Supply Chain
Zahra Mobini, Erasmus University Rotterdam, Rotterdam,
Netherlands, mobinidehkordi@ese.eur.nl, Albert Wagelmans,
Wilco Van Den Heuvel
48-Room 105A, CC
Managing Contracts and Financial Flow in
Supply Chain
Sponsor: Manufacturing & Service Oper Mgmt/iFORM
Sponsored Session
In a two-echelon supply chain consisting of a supplier and a retailer where the
latter has private information about his cost parameter, we analyze the design of
the supplier’s optimal menu of contracts. Instead of offering a separating menu,
the supplier offers a menu of bunching contracts where each contract is intended
to appeal to more than one retailer type. We investigate the effects of offering
such a menu on the supplier’s profit.
Chair: Lingxiu Dong, Associate Professor, Washington
University in St. Louis, One Brookings Drive, St. Louis, MO, 63132,
United States of America, dong@wustl.edu
1 - Buyer Intermediation in Supplier Finance
Tunay Tunca, ttunca@rhsmith.umd.edu, Weiming Zhu
4 - Overstock Goods Auctions
Zohar Strinka, PhD Candidate, University of Michigan,
1205 Beal Ave., Ann Arbor, MI, 48109, United States of America,
zstrinka@umich.edu, H. Edwin Romeijn
We analyze the role and the efficiency of buyer intermediation in supplier
financing (BIF). We theoretically demonstrate that BIF can significantly improve
the supply chain surplus over traditional financing. Using data from a large
Chinese online retailer, we estimate model parameters, empirically verify the
theory, and predict efficiency gains.
Manufacturers sometimes find themselves with a considerable quantity of
overstock goods. In these cases, some turn to online liquidation auctions to sell
excess inventory. We propose implementing US Treasury-style auctions in this
setting which allow bidders to specify pairs of bid price and a desired quantity at
that price. We consider bidders who are themselves retailers and face
newsvendor-type costs based on the number of units won and uncertain demand.
2 - Financial Pooling in Supply Chains
S. Alex Yang, Assistant Professor, London Business School, Sussex
Place, London, United Kingdom, sayang@london.edu, Qu Qian,
Ming Hu
Trade credit pools liquidity between suppliers and retailers. Due to this pooling
effect, even if the supplier’s cost of capital is higher, the retailer may still demand
for trade credit. Supply chain finance increases the efficiency of this pooling
effect, and hence reduces the overall chain financing cost.
■ TA50
50-Room 106A, CC
3 - Trade Credit and Supplier Competition
Jiri Chod, Boston College, Carroll School of Management,
Chestnut Hill, MA, jiri.chod@bc.edu, S. Alex Yang,
Evgeny Lyandres
Operations Management and Marketing Interface
We study the effect of competition among suppliers on their willingness to
provide trade credit. Providing trade credit to a financially constrained buyer
allows this buyer to reallocate his cash budget to purchasing from competing
suppliers. Thus, relaxing the buyer’s financial constraint may backfire at the
supplier who provides financing. This is a possible explanation of the empirical
regularity that firms selling differentiated products tend to offer more trade credit.
Chair: Ozge Sahin, Johns Hopkins University, ozge.sahin@jhu.edu
Sponsor: Manufacturing & Service Operations Management
Sponsored Session
Co-Chair: Yao Cui, Cornell University, 401N Sage Hall, Ithaca, United
States of America, yao.cui@cornell.edu
1 - Econometric Models of Pairwise Externalities and Social
Attractiveness for the Music Industry
Stefano Nasini, Post-doctoral Researcher, IESE Business School,
3-7, Arnus i Gari, Barcelona, Spain, snasini@iese.edu,
Victor Martínez-de-Albéniz
4 - Push, Pull, and Delayed Payment Contracts when a Manufacturer
Expands His Product Line
Xiaomeng Guo, PhD Candidate, Olin Business School,
Washington University in St. Louis, Campus Box 1156, One
Brookings Drive, St. Louis, MO, 63130, United States of America,
xiaomeng.guo@wustl.edu, Lingxiu Dong, Danko Turcic
We developed an econometric model of social attractiveness that integrates time
variation of individual decisions with the structural information concerning their
spillovers. The exponential family of distributions is used to jointly deal with the
dynamic and structural aspect of such a complex statistical setting. It resulted in a
well-suited model for the analysis of artist goods. An application to a large data
set of song diffusion on the radio is presented.
A manufacturer’s ability to sell a new product often depends on a retailer’s
willingness to stock the product. We construct a game-theoretic model of a supply
chain with stochastic, price-sensitive demand and consider three basic wholesale
price contracts: push, pull and delayed payment contracts. We show how a
manufacturer can influence the retailer’s incentive to carry a second product by
choosing a “correct” contract type and clarify which contract should be expected
in equilibrium.
2 - Inventory Management for Luxury Goods
Ruslan Momot, INSEAD, Boulevard de Constance,
Fontainebleau, 77305, France, ruslan.momot@insead.edu,
Elena Belavina, Karan Girotra
Firms selling conspicuous goods face a trade-off: producing more allows for
extracting more revenues but compromises the product’s reputation for
exclusivity. We capture this trade-off in a dynamic model of strategic customer
and firm behavior that includes limited memory. Firms should follow stationary
cyclic strategies alternating scarcity and overproduction. The former builds a
reputation whereas the latter exploits it. The longer the customer memory,
shorter is the overproduction phase.
■ TA49
49-Room 105B, CC
Uncertainty in Sourcing and Procurement
Sponsor: Manufacturing & Service Oper Mgmt/Supply Chain
Sponsored Session
3 - A Newsvendor Model with Product Bundling
Qingning Cao, University of Science and Technology of China, 96
Jinzhai Rd, SM 611, Hefei, China, caoq@ustc.edu.cn, Jun Zhang,
Kathy Stecke, Xianjun Geng
Chair: Zohar Strinka, PhD Candidate, University of Michigan,
1205 Beal Ave., Ann Arbor, MI, 48109, United States of America,
zstrinka@umich.edu
1 - Supplier Diversification under Random Yield and Price
Dependent Demand
Guang Xiao, Olin Business School, Washington University in St.
Louis, St. Louis, United States of America, xiaoguang@wustl.edu,
Lingxiu Dong, Nan Yang
This paper studies a firm’s optimal ordering decision of a primary product when
the firm can bundle this product with another product. The firm makes an
ordering decision before demand uncertainty resolves, and then retails this
primary product either alone or in a mixed bundle with a secondary product. Our
results suggest that as compared to a no-bundling benchmark, the firm should
overstock (understock) when the wholesale price is high (low).
274
4 - Strategic Consumers, Revenue Management and the Design of
Loyalty Programs
So Yeon Chun, McDonough School of Business, Georgetown
University, 3700 O St. NW, Washington, DC, United States of
America, sc1286@georgetown.edu, Anton Ovchinnikov
TA53
2 - Price Drop Protection Policy with Partial Refunds
Dinah Cohen-Vernik, Assistant Professor Of Marketing,
Rice University, 6100 Main St, Houston, TX, 77006,
United States of America, dv6@rice.edu, Amit Pazgal
Many retailers now offer to refund customers the full price difference as long as
the price drop occurred within a specified short period of time after the purchase.
Despite the popularity of such policy, the existing marketing research on the topic
is scarce. In this paper we investigate the price difference refund policy (referred
to as price drop protection) and demonstrate how it can improve retailer’s profits.
Several major airlines recently switched their loyalty programs from
``mileage/segment-based” toward ``spending-based”. We study the impact of this
switch on firm’s profit and consumer utility. We present a novel model of strategic
consumers’ response to firm’s pricing and loyalty program decisions, incorporate
such response into the firm’s pricing and loyalty program design problem,
compare the solutions under the mileage-based versus spending-based design,
and discuss managerial implications.
3 - Clicks and Editorial Decisions: How Does Popularity Shape Online
News Coverage?
Pinar Yildirim, Assistant Professor Of Marketing, The Wharton
School. University of Pennsylvania, 748 Huntsman Hall,
Philadelphia, PA, 19104, United States of America,
pyild@wharton.upenn.edu, Ananya Sen
■ TA51
Using online news data from a large Indian English daily newspaper, this paper
analyzes how demand side incentives shape news media reporting. To establish a
causal link, we instrument the views of articles using days with rain and days
with electricity shortage as exogenous shocks to reader attention. We provide
evidence for extended coverage and higher resource allocation to issues which
receive high number of clicks.
51-Room 106B, CC
Economics of Innovation in Supply Chains
Sponsored Session
Chair: Ayhan Aydin, Assistant Professor Of Operations Management,
George Mason University School of Business, 4400 University Drive MS
5F4, Fairfax, VA, 22030, United States of America, aaydin2@gmu.edu
1 - Product Quality in a Decentralized Supply Chain: Value of
Information Asymmetry
Narendra Singh, Narendra.Singh@scheller.gatech.edu,
Stylianos Kavadias, Ravi Subramanian
4 - Stock-out Detection System Based on Sales Transaction Data
Ricardo Montoya, Assistant Professor, University of Chile,
Republica 701, Santiago, Chile, rmontoya@dii.uchile.cl,
Andres Musalem, Marcelo Olivares
We present a methodology based on real-time point-of-sales data to infer on-shelf
product availability. We develop our methodology using process control theory an
apply it to a big-box retailer. We use historical transactional data to develop our
methodology and empirically test it in two field studies. We analyze the results
and implications.
We study an OEM’s optimal product design quality and sourcing strategies in a
supply chain consisting of an OEM, who has in-house option, and a supplier, who
has more favorable cost structure and the power to dictate contract terms. We
show that a two-part tariff contract, as opposed to a price-only contract, may
leave both the OEM and the supplier worse off. Further, we show that
asymmetric information about the OEM’s cost structure may lead to higher profits
for both the OEM and the supplier.
■ TA53
53-Room 107B, CC
Behavior in Operational Contexts
2 - Information Acquisition and Innovation in Competitive Markets
Yi Xu, Associate Professor, Smith School of Business, University
of Maryland, College Park, MD, 20742, United States of America,
yxu@rhsmith.umd.edu, He Chen, Manu Goyal
Sponsor: Behavioral Operations Management
Sponsored Session
In this paper, we study firms’ information acquisition strategies and innovation
strategies in a competitive market with uncertainty. The firms can resolve the
market uncertainty through different information acquisition methods. We
highlight the strategic interactions between information acquisition and
innovation investments in such a market.
Chair: Anton Ovchinnikov, Queen’s University, 143 Union Str, West,
Kingston, Canada, anton.ovchinnikov@queensu.ca
1 - Behavioral Ordering: Inventory, Competition and Policy
Bernardo Quiroga, Assistant Professor, Business And Behavioral
Science, Clemson University, 100 Sirrine Hall, Clemson, SC,
29634, United States of America, bfquirog@gmail.com,
Anton Ovchinnikov, Brent Moritz
3 - Investment in Core Technologies and Consumer Markets
Ayhan Aydin, Assistant Professor Of Operations Management,
George Mason University School of Business, 4400 University
Drive MS 5F4, Fairfax, VA, 22030, United States of America,
aaydin2@gmu.edu, Rodney Parker
We study the effect of observed inventory decisions on performance. Our goal is
to measure and understand profit losses due to behavioral (intuitive but
suboptimal) ordering. The current literature, primarily focused on a newsvendor
making decisions in isolation, reports results implying profit losses of 1-5%
compared to the analytical optimum. In contrast, we show that when a
behavioral inventory manager competes against a management-science-driven
competitor, profit losses are much larger.
We consider a two-tier supply chain, an upstream tier composed of two
competing providers of a component that is used by multiple OEMS (integrators)
in the lower tier. Upstream firms invest to develop the technology of the
component further. We investigate the effects of downstream market factors, the
nature of technology, competition, and the level of uncertainties in the R&D
process on the level of upstream investments and the adoption of the higher
technologies by the downstream firms.
2 - Inequity and Loss Aversion in Pay What You Want
Yulia Vorotyntseva, PhD Candidate, The University of Texas at
Dallas, Richardson, United States of America,
Yulia.Vorotyntseva@utdallas.edu, Ozalp Ozer
Pay-What-You-Want pricing is an exemplar of fairness-driven behavior in a
business context: the price for a product is fully determined by a buyer, and the
seller cannot reject any offer. The objective of our work is to find out key factors
affecting the buyers’ selection of prices under PWYW. We use a distributional
fairness approach and build a hierarchical Bayesian model of buyers’ behavior.
We then test it in a controlled laboratory experiment.
■ TA52
52-Room 107A, CC
Consumer-driven Management Science
Sponsor: Marketing Science
Sponsored Session
3 - Inventory Decisions in the Presence of Strategic Consumers
Yaozhong Wu, National University of Singapore, NUS Business
School, Singapore, Singapore, yaozhong.wu@nus.edu.sg,
Yang Zhang, Benny Mantin
Chair: Ricardo Montoya, Assistant Professor, University of Chile,
Republica 701, Santiago, Chile, rmontoya@dii.uchile.cl
1 - Product Showcasing in the Presence of Experience Attributes
Daria Dzyabura, Assistant Professor of Marketing, NYU Stern
School of Business, 40 West 4th Street, Tisch 805, New York, NY,
10012, United States of America, ddzyabur@stern.nyu.edu,
Srikanth Jagabathula
In the presence of strategic consumers, who may delay their purchase to the
markdown season, a retailer is faced with an extra consideration in addition to
the traditional newsvendor setting: excess inventory may induce strategic
consumers to delay their purchase and may further harm the revenue. We
develop a model that accounts for both the strategic consumers and the retailer’s
inventory decisions. We design behavioral experiments to test our model
predictions.
We formalize a firm’s showcase decision, or selecting a subset of products to carry
in a physical store, while a ‘large’ product line is offered through the online
channel. Some customers visit the offline store to gain information about product
features. We formalize the showcase problem as an IP, which we show to be NPcomplete, derive closed-form solutions for special cases, and adapt the local search
heuristic to the general problem. We gather conjoint data to estimate the model
parameters.
275
TA54
4 - When to Hire the First Employee? Behavioral Evidence
and Insights
Beatrice Boulu-reshef, Behavioral Research Associate, Darden
School of Business, 100 Darden Boulevard,
Charlottesville, VA, 22903, United States of America,
Boulu-ReshefB@darden.virginia.edu, Anton Ovchinnikov,
Charles Corbett
3 - Slacks-based Measure Variations Revisited
Kaoru Tone, Professor, National Graduate Inst. for Policy Studies,
7-22-1 Roppongi, Minato-ku, Tokyo, 106-8677, Japan,
tone@grips.ac.jp
In Tone (2010), I developed four variants of the SBM model where main concerns
are to search the nearest point on the efficient frontiers of the production
possibility set. However, in the worst case, a massive enumeration of facets of
polyhedron associated with the production possibility set is required. In this
paper, I will present a new scheme for this purpose which requires a limited
number of additional linear program solutions for each inefficient DMU.
Effectively any entrepreneur shifts from doing all the work him/herself to hiring
someone to do part of that work. We use an analytical model and behavioral
experiments to study when entrepreneurs should and do hire their first
employee. Understanding both the optimal timing/conditions of hiring and the
deviations of the hiring patterns from optima have the potential to provide
insights to a very broad spectrum of entrepreneurs at the critical early stage of
their new venture formation process.
■ TA56
56-Room 109A, CC
Execution Mode Choices for NPD
■ TA54
Cluster: New Product Development
Invited Session
54-Room 108A, CC
Applying Machine Learning in Online Revenue
Management
Chair: Pascale Crama, Singapore Management University, 50 Stamford
Road, Singapore, 178899, Singapore, pcrama@smu.edu.sg
1 - Managing Exploration and Execution
Nittala Lakshminarayana, University of California San Diego,
9256 Regents Road Apt. G, La Jolla, CA, 92037, United States of
America, Lakshminarayana.Nittala@rady.ucsd.edu, Sanjiv Erat,
Vish Krishnan
Cluster: Tutorials
Invited Session
Chair: David Simchi-Levi, Professor, Massachusetts Institute of
Technology, 77 Massachusetts Avenue, Cambridge, MA, 02139,
United States of America, dslevi@mit.edu
1 - Tutorial: Applying Machine Learning in Online
Revenue Management
David Simchi-Levi, Professor, Massachusetts Institute of
United States of America, dslevi@mit.edu
We model Innovation as a multi-stage activity consisting of Exploration and
Execution. Within this parsimonious model that mimics many contexts in
Innovation, we consider the effect of incentives and several institutional features
on the optimal idea generation and development strategy.
2 - Customer Co-design: The Role of Product Lines
Sreekumar Bhaskaran, sbhaskar@mail.cox.smu.edu, Amit Basu
In a dynamic pricing problem where the demand function is unknown a priori,
price experimentation can be used for demand learning. In practice, however,
online sellers are faced with a few business constraints, including the inability to
conduct extensive experimentation, limited inventory and high demand
uncertainty. In this talk we discuss models and algorithms that combine machine
learning and price optimization that significantly improve revenue. We report
results from live implementations at companies such as Rue La La, Groupon and a
large European Airline carrier.
Involving customers in the new product design can be a powerful means to
achieve high levels of customer satisfaction and market success. However, the
“co-design” process may require participating customers to commit significant
time and effort, while facing the uncertainty that the firm may overprice the
custom product. Since this reduces a customers incentive to commit effort upfront, co-design can be difficult to motivate. We develop analytical models that
capture these various effects.
3 - Flexibility and Knowledge Development in Product Development:
Insights from a Landscape Search Model
Mohsen Jafari Songhori, Jsps Research Fellow, Tokyo Institute of
Technology, J2 Bldg., Room 1704, 4259 Nagatsuta-cho,, Tokyo,
226-8502, Japan, mj2417@gmail.com, Majid Abdi, Takao Terano
■ TA55
55-Room 108B, CC
This study introduces a landscape model of Product Development (PD). The
model captures different PD performance aspects (e.g. development time, quality
and cost) and their trade-offs. Moreover, knowledge development dynamics and
flexibility are incorporated in the model to investigate how strategies toward
these, in PD process, are associated with the performance measures.
Extensions of DEA
Cluster: Data Envelopment Analysis
Invited Session
Chair: Endre Bjorndal, Associate Professor, Norwegian School of
Economics, Helleveien 30, Bergen, 5045, Norway,
Endre.Bjorndal@nhh.no
1 - Assessment of Alternative Approaches to Include Exogenous
Variables in DEA Estimates
Jose M. Cordero, Universidad de Extremadura, Av Elvas sn,
Badajoz, Spain, jmcordero@unex.es, Daniel Santin
■ TA57
57-Room 109B, CC
Applications of Stochastic and Dynamic
Programming in Energy
The aim of this paper is to compare the performance of some recent methods
developed in the literature to incorporate the effect of external variables into the
estimation of efficiency measures such as the conditional approach developed by
Daraio and Simar (2005, 2007) or the one-stage model proposed by Johnson and
Kuosmanen (2012). To do this, we conduct a Monte Carlo experiment using a
translog function to generate simulated data.
Sponsor: ENRE – Energy II – Other (e.g., Policy, Natural Gas,
Climate Change)
Sponsored Session
Chair: Andrew Liu, Assistant Professor, Purdue University,
315 N. Grant Street, West Lafayette, IN, 47907,
United States of America, andrewliu@purdue.edu
1 - Approximate Dynamic Programming for Pricing-based Real-time
Demand Management
Ozgur Dalkilic, The Ohio State University, 205 Dreese Labs, 2015
Neil Ave, Columbus, OH, 43210, United States of America,
dalkilic.1@buckeyemail.osu.edu, Atilla Eryilmaz, Antonio Conejo
2 - Compensating for Exogenous Cost Drivers in the Regulation of
Electricity Networks
Endre Bjorndal, Associate Professor, Norwegian School of
Economics, Helleveien 30, Bergen, 5045, Norway,
Endre.Bjorndal@nhh.no, Maria Nieswand, Mette Bjørndal,
Astrid Cullmann
The present yardstick model used by the Norwegian regulator compensates, via
two-stage DEA efficiency analysis, for a number of environmental factors. These
factors are correlated with measured efficiency and company size. We compare
conditional nonpararametric methods to current benchmarking model, and we
discuss whether the choice of method affects the revenue caps of companies in a
systematic manner.
We consider the real-time demand management problem of a load aggregator that
coordinates the consumer demand to match a predetermined daily load. The
aggregator’s objective is to minimize its payment to the real-time market. Under
uncertainty of the market prices, we derive dynamic pricing algorithms that
approximate the optimal dynamic programming solution. We show via numerical
investigations that the proposed algorithms coordinate flexible demand and
achieve close to optimal allocation.
276
TA60
2 - A Revenue Adequate Stochastic Programming Market
Clearing Mechanism for Effective Integration of Volatile
Renewable Generation
Golbon Zakeri, Dr., University of Auckland,
Auckland, New Zealand, g.zakeri@auckland.ac.nz
4 - Volumes for the Renewable Fuel Standard using Multiobjective
Programs with Equilibrium Constraints
Sauleh Siddiqui, Assistant Professor, Johns Hopkins University,
3400 N. Charles St. Latrobe 205, Baltimore, MD, 21218, United
States of America, siddiqui@jhu.edu, Adam Christensen
Pentration of generation from volatile renewable sources of electricity generation
has substantially increased in electricity markets around the world. Various
authors (e.g. Pritchard et al and Morales et al) have proposed mechanisms to deal
with this however these approaches are not revenue adequate under each
scenario (although they are in expectation). We will introduce a variation of the
stochastic dispatch mechanism that is revenue adequate under each scenario and
present its properties.
We apply a Multiobjective Program with Equilibrium Constraints to the United
States renewable fuel market to help understand why it has been so difficult in
releasing the 2014 mandate for the Renewable Fuel Standard. Our analysis
provides a variety of policy alternatives to aid in setting these volume obligations
and is applicable to a wide variety of climate and energy market settings.
3 - Parallel Computing of Stochastic Programs with Application to
Energy System Capacity Expansion
Andrew Liu, Assistant Professor, Purdue University, 315 N. Grant
Street, West Lafayette, IN, 47907, United States of America,
andrewliu@purdue.edu, Run Chen
■ TA59
59-Room 110B, CC
Fire Management 1: Suppression
Sponsor: ENRE – Environment II – Forestry
Sponsored Session
Power grids’ planning and operations exhibit extreme multiscale, ranging from
hourly operation to decades of planning. The linkage between decisions at
different time scales may be relaxed to produce multiple independent
subproblems. We propose to use an augmented Lagrangian multiplier method to
design parallel algorithms to solve such multiscale problems. Convergence of the
embedded algorithm for convex problems will be shown, along with preliminary
numerical results.
Chair: Vitaliy Krasko, Colorado School of Mines, 1500 Illinois St,
Golden, CO, United States of America, vkrasko@mymail.mines.edu
1 - Efficient use of Aerial Firefighting Assets
Matthew Thompson, US Forest Service, 800 E Beckwith,
Missoula, MT, 59801, United States of America,
mpthompson02@fs.fed.us
■ TA58
This presentation will explore themes in measuring and improving the efficiency
of aerial firefighting assets. Insights from case studies in the US and Italy will be
highlighted.
58-Room 110A, CC
Topics in Oil, Natural Gas, and Alternative Fuels
2 - The Development and use of Forest Fire Detection System
Performance Measures
Dave Martell, david.martell@utoronto.ca
Climate Change)
Sponsored Session
Forest fire detection systems are designed to detect fires while they are small but
concerted efforts to minimize detection size can enhance the performance of the
detection sub-system at the expense of overall fire management system
performance. We describe the development and use of a detection system
performance measure designed to overcome such problems.
Chair: Sauleh Siddiqui, Assistant Professor, Johns Hopkins University,
3400 N. Charles St. Latrobe 205, Baltimore, MD, 21218,
United States of America, siddiqui@jhu.edu
1 - Rate-of-return Regulation and Investment in a
Natural Gas Pipeline
Olivier Massol, IFP School, 228-232 avenue Napoléon Bonaparte,
Rueil-Malmaison, France, olivier.massol@ifpen.fr,
Florian Perrotton
3 - A Network Interdiction Approach for Mitigating a
Pyro-Terror Attack
Eghbal Rashidi, PhD Student, Mississippi State University,
1212 Louisville St, # 58, Starkville, MS, 39759,
United States of America, er442@msstate.edu, Hugh Medal
We study a problem in which a group of terrorists seek to maximize the impact of
a pyro-terror attack by optimally locating the fire ignition points on a landscape,
whereas fire managers wish to interdict the expansion of fire and mitigate the
damage using an optimal fuel treatment plan. We model the problem as a
Stackelberg game and develop a decomposition algorithm to solve it.
We examine the economics of a natural gas pipeline project provided by a foreign
private firm in a LDC. The infrastructure could trigger possible future
developments of the domestic gas sector. We address two questions. First, can a
rate-of-return (ROR) type of regulatory organization induce the firm to rationally
accept to build ahead of proven demand? Second, how far would the allowed
ROR have to rise for the pipeline design to be adopted be congruent with the
socially desirable one?
4 - Machine Learning Methods to Improve Fire Suppression Policies
on Simulated Landscapes
Hailey Buckingham, hailey.buckingham@oregonstate.edu,
Claire Montgomery
2 - Multistage Stochastic Model for Natural Gas Contract and
Maintenance Scheduling of Power Plants
Zhouchun Huang, University of Central Florida, 4000 Central
Florida Blvd, Orlando, Fl, 32816, United States of America,
hzclinger@gmail.com, Qipeng Zheng
Any policy which informs wildfire suppression decisions affects the future
evolution of the landscape as patterns of growth, harvest, and fire each adjust to
the influence of the policy’s fire suppression regime. Improving a policy is difficult
because the long-term effects may not be obvious a priori because present
decisions change future states. In this study, we use machine learning techniques
and monte carlo simulations of forested landscapes to improve a wildfire
suppression policy.
Natural gas contracting and equipment maintenance scheduling are two major
factors that affect the profit of a natural gas power plant. We consider both of
them together and propose a multistage stochastic model to address the
uncertainties of gas and electricity prices in the market. A scenario-based
decomposition strategy is applied to solve the model and the numerical results
will be present.
■ TA60
3 - A Crude Oil Market Model for the United States
Olufolajimi Oke, PhD Candidate, Johns Hopkins University, 3400
N Charles St, Latrobe 205, Baltimore, MD, 21218, United States
of America, ooke1@jhu.edu, Max Marshall, Ricky Poulton,
Sauleh Siddiqui, Daniel Huppmann
60-Room 111A, CC
Education I
Contributed Session
The United States’ crude oil industry currently faces infrastructural and
environmental challenges, as production surges and crude-by-rail shipments and
incidents increase. We adapt Huppman and Egging’s dynamic Generalized Nash
Equilibrium model (Multimod) to the US market. Thus, we can analyze the
transportation of crude oil and explore possible scenarios to recommend decisions
for safe movement, as well as gauge the economic impact of new regulations and
policy interventions.
Chair: Nabil Belacel, Senior Research Officer, NRC, 100 des Aboiteaux
Street, Moncton, NB, E1A7R1, Canada, nabil.belacel@nrc.gc.ca
1 - Co-Author Network Analysis of Operations Management Journals
Bonie(he) Zhang, PhD Candidate, Rutgers Business School,
1 Washington Park, Newark, United States of America,
boni.zhang@rutgers.edu, Yao Zhao, Xinxin Xuan
We study the co-author network of flag-ship INFORMS journals in operations
management such as Management Science and Operations Research. Our
empirical exploration characterizes the changing patterns of the co-author
network and provides insights to authors on how to improve productivity
through exploitation of the academic social network.
277
TA61
2 - Case Study: Vastrapur Car Rental Services
Balaraman Rajan, Assistant Professor, California State University
East Bay, 25800 Carlos Bee Blvd, Hayward, CA, 94542,
United States of America, balaraman.rajan@csueastbay.edu,
Ravichandran Narasimhan
3 - Equilibrium Investment Strategies in Renewable Portfolio
Standards under Uncertainty
Yuta Kamobayashi, Tokyo University of Science, 2641 Yamazaki,
Noda-shi, Chiba, Japan, 7414609@ed.tus.ac.jp, Ryuta Takashima,
Makoto Tanaka, Yihsu Chen
In this case we discuss the revenue model for a rental car business in India. The
case can be used for teaching topics in probability and decision modeling at both
undergraduate and graduate level. The first part of the case focuses on expected
value and the second part of the case involves decision making under uncertainty
and strategic choices. It can also be extended to train students in basic simulation
using Crystal Ball or other such tools.
Recently renewable portfolio standard (RPS) has been introduced due to further
penetration of renewable energies. In this paper, we propose a two-period
competition model in an oligopolistic electricity industry with uncertain demand
in order to consider investment behaviors for firms in a framework of the PRS.
We analyze an effect of the RPS on investments in renewables and nonrenewables. Additionally, we show how a percentage of production from
renewables affects the market equilibrium.
3 - A Learner-Analytics Based Approach for Attenuating the
Course-Level Dropout Rate
Aysegul Demirtas, Graduate Student, Arizona State University,
699 S Mill Avenue, Tempe, AZ, 85281, United States of America,
ademirt2@asu.edu, Jennifer Bekki, Esma Gel, George Runger
4 - Analysis of Regional Market Impact of EPA’s Clean Power Plan:
Mass-based or Rate-based Standard?
Duan Zhang, University of California, Merced, 1392 Dynes St,
Merced, CA, 95348, United States of America,
dzhang8@ucmerced.edu, Yihsu Chen, Makoto Tanaka
Despite their potential to attract larger numbers of students, online courses
remain plagued by a student attrition problem. We apply data mining and learner
analytics techniques to better understand online learner behavior in an effort to
attenuate the online course drop-out rate. We present our modeling approach,
utilizing data from student interactions with the course LMS, and our findings on
course-level persistence based on the application of our approach to data from
multiple courses.
We studied the market and emission outcomes of the EPA proposed rate-based
emission policy or the Clean Power Plan. A theoretical model was built to
generate contestable hypothesis. A large-scale simulation of the PennsylvaniaJersey-Maryland electricity market in 2012 was used to validate the hypotheses
and quantify the magnitude of impacts, including distribution of economics rent
as well as the shift of pollution emissions. We report the preliminary results in
this talk.
4 - The School Closing Problem
Jing Xu, University of Pennsylvania, 209S 33rd Street,
Department of Mathematics, Philadelphia, PA, 19104,
United States of America, xjing@sas.upenn.edu
■ TA62
When school districts face declining enrollments, schools must be closed to reduce
costs. The choice of which schools to be shuttered is controversial. Surprisingly,
few papers have considered this problem. This paper considers the effect of using
existing school choice mechanisms to close schools. It turns out simple
modifications of existing algorithms produce perverse results. We also establish
non-existence of a Pareto-efficient and strategy-proof mechanism in a basic
school closing model.
62-Room 112A, CC
Reliability and Random Factors in Power Systems
Cluster: Energy Systems: Design, Operation, Reliability and
Maintenance
Invited Session
Chair: Bo Zeng, Assistant Professor, University of South Florida, Tampa,
4202 E. Fowler Avenue, Tampa, Fl, 33620, United States of America,
bzeng@usf.edu
1 - Protect Power System from Electromagnetic Pulse
Feng Pan, Research Engineer, Pacific Northwest National
Laboratory, P.O. Box 999 MSIN K1-85, Richland, WA, 99352,
United States of America, feng.pan@pnnl.gov, Russell Bent,
Aric Hagberg
■ TA61
61-Room 111B, CC
Sustainability in Energy Sector: Policy Analysis and
Technology Assessment
Sponsor: ENRE – Environment I – Environment and Sustainability
Sponsored Session
Power grids are vulnerable to Electromagnetic pulse (EMP) that can lead a power
grid to collapse in a short time. We introduce an optimization model to configure
a power grid prior to an EMP so that the damage caused by EMP is reduced. This
talk will focus on the modeling aspect.
Chair: Yihsu Chen, Associate Professor, University of California,
Merced, 5200 N. Lake Rd, Merced, CA, 95343,
United States of America, ychen26@ucmerced.edu
1 - Market Impacts of Energy Storage in a Transmission-Constrained
Power System
Afzal Siddiqui, University College London, Department of
Statistical Science, Gower Street, London, UK, WC1E 6BT, United
Kingdom, afzal.siddiqui@ucl.ac.uk, Vilma Virasjoki, Paula Rocha,
Ahti Salo
2 - Modeling Cascading Failures and Restoration Times in Power
Networks to Address Resilience
Sinan Tas, Assistant Professor, Penn State University-Berks
College, sut12@psu.edu, Vicki Bier
Prevention is generally the default solution in security investments of critical
infrastructure. Electric power networks are capacity-constrained systems, which
makes them a perfect candidate for cascading failure. Moreover, different
components take substantially different times to recover. In this study, we will
analyze investments that will possibly improve overall resilience of the network
(rather than preventive ones that decreases the likelihood of such attacks).
Intermittent renewable energy (RE) technologies require conventional power
plants to ramp up more often. In turn, energy storage may offset the
intermittency of RE technologies and facilitate their integration into the grid. In
order to assess the consequences of storage, we use a complementarity model
with market power, transmission constraints, and uncertainty in RE output. We
find that although storage reduces congestion and ramping costs, it may actually
increase greenhouse gas emissions.
3 - Joint Planning of Energy Storage and Transmission for Wind
Energy Generation
Wei Qi, PhD Candidate, University of California, Berkeley, 1117
Etcheverry Hall, Berkeley, CA, 94720, United States of America,
qiwei@berkeley.edu, Yong Liang, Zuo-jun Max Shen
2 - Do Emissions Caps Lead to Carbon Leakage in Regional
Markets? The Case of South-east Europe
Verena Viskovic, PhD Student, University College London,
50 Tiber Gardens, London, N/, N10XE, United Kingdom,
verena.viskovic@gmail.com, Yihsu Chen, Afzal Siddiqui
Abstract: Regions with abundant wind energy usually have no ready access to
power infrastructure. We propose models of transmission network planning with
co-location of energy storage systems for wind energy delivery. Our models
determine the sizes and sites of storage stations as well as the corresponding
topology and capacity of the transmission network. Then we present various
insights regarding storage value, technology advancements and layout robustness.
We examine the extent of carbon leakage in neighbouring jurisdictions with
different carbon emissions reduction policies. We use a complementarity model to
illustrate carbon leakage on a three-node power system. Subsequently, we model
a 19-node Southeastern European network in order to study carbon leakage on
the periphery of the EU.
278
■ TA63
TA65
2 - Lessons Learned Deploying ODG
Larry Neal, Independent, 3667 Cantelow Rd, Vacaville, CA,
95688, United States of America, lnealjr@wildblue.net,
Frank Koch
63-Room 112B, CC
KINFORMS Sponsored Session
A panel of seasoned practitioners will discuss the lessons learned in deploying the
concepts of Decision Quality throughout their organization, or ODQ. After brief
opening remarks, the panel will discuss the learnings both positive and negative,
of their experiences. The focus of this session is to help other institutions follow
suit and raise the bar on their organization decision making practices. Attendees
will come away with readily usable insights and tips for their own use.
Sponsor: KINFORMS
Sponsored Session
Chair: Chang Won Lee, Corresponding Author, Hanyang University,
School of Business, Seoul, 133-791, Korea, Republic of,
leecw@hanyang.ac.kr
1 - Study on the Supply Chain Management Critical Success
Factors (csf)
Chang Won Lee, Corresponding Author, Hanyang University,
School of Business, Seoul, 133-791, Korea, Republic of,
leecw@hanyang.ac.kr, Gary Gaukler
3 - Applying Decision Analysis at Pfizer – Lessons Learned from
the Field
Rodger Thompson, Sr. Director/team Leader, Pfizer, Inc., 500
Arcola Road, Collegeville, PA, 19426, United States of America,
rodger.thompson@pfizer.com
This presentation will discuss the journey that the Portfolio and Decision Analysis
(PDA) group at Pfizer has undertaken to bring decision excellence to the Pfizer
organization. The discussion will focus on lessons learned on adapting the Dialog
Decision Process to Pfizer to enable integration of the six components of decision
quality.
For many companies, managing their supply chain has become increasingly
central to their business success. Thus, it is crucial to investigate and identify
appropriate supply chain practices for today’s business environment. We call these
practices, the Critical Success Factors (CSF) for supply chain management (SCM).
Appropriate measures are developed and tested with a questionnaire survey. The
results of the empirical analysis confirm that SCM-CSF can be conceptualized.
2 - Retailer’s Optimal Sourcing Strategy under Consumer Stockpiling:
A Risk Management Approach
Jiho Yoon, Michigan State University, N468 North Business
Complex, Michigan State University, East Lansing, MI, 488241121, United States of America, yoon@broad.msu.edu,
Ram Narasimhan, Myungkyo Kim
■ TA65
We study a retailer’s sourcing strategy under consumers’ stockpiling behavior and
the factors associated with the selection of an optimal strategy in multi-tier supply
chains in the presence of supply disruption risk. Stockpiling behavior occurs
when consumers attempt to mitigate the negative impact of a supply shortage.
Our analysis shows that optimal sourcing strategy is highly dependent on multiple
factors.
Sponsor: Decision Analysis
Sponsored Session
65-Room 113B, CC
Recent Findings and Experiences in
Probability Elicitation
Chair: Saurabh Bansal, Assistant Professor, Penn State Univrsity, 405
Business Building, University Park, PA, 16802, United States of
America, sub32@psu.edu
1 - Indirect Elicitation of Subjective Probabilities through
Pair-Wise Comparisons
David Budescu, Professor, Fordham University, 441 E Fordham
Road, 220 Dealy Hall, Bronx, NY, 10458, United States of
America, budescu@fordham.edu, Han Hui Por
3 - Relationships in Servitization, Satisfaction and Intention to Reuse:
Customers’ Perspective
Sang Hyung Ahn, Professor, Seoul National University, Graduate
School of Business, Seoul, Korea, Republic of, shahn@snu.ac.kr,
Chang Won Lee
This study presents to find out a relationship among characteristics of
servitization, satisfaction and intention to reuse in terms of customers’
perspective. The results were examined to identify significant factors affecting
servitization, satisfaction and intention to reuse. The study provides decisionmakers with more accurate information to develop appropriate servitization
practices in terms of customers perspective.
We test a new method for eliciting subjective probabilities. Judges compare pairs
of possible outcomes and identify which of the two is more likely, and by how
much. These judgments generate a matrix from which the target probabilities are
estimated by the geometric means. We compared the quality of our estimates
with traditional direct estimates and show that they were significantly more
accurate, suggesting that the new approach is a good candidate for replacing
standard elicitation methods.
4 - Industrialization, Productivity and the Shift to Services
and Information
Hosun Rhim, Professor Of Logistics, Service, And Operations
Management, Korea University Business School, Anam-dong,
Seongbuk-gu, 136-701, Seoul, Korea, Republic of,
hrhim@korea.ac.kr, Uday Karmarkar, Kihoon Kim
2 - Eliciting and Modeling Continuous Forecasts
Joe Tidwell, University of Maryland, Biology/Psychology
Building, College Park, United States of America,
jtidwell@umd.edu
Accurate forecasting models for continuous outcomes offer many benefits,
including eliminating most close-call counterfactuals, better information about
tail risks, and the ability to obtain forecasts for any value across the range of
possible outcomes. In a series of experiments, we evaluate various methods for
eliciting small sets of judgments from individual forecasters regarding real-world
events and then aggregating these judgments over forecasters into continuous
forecast models.
The traditional explanation for the shift to services was the steady growth of
manufacturing productivity. But this does not explain the initial growth in
manufacturing, or that of information intensive services relative to physical
services. The authors adduce a second factor that explains both trends: the
relative maturity of a market.
3 - Estimating Continuous Distributions by Quantifying Errors in
Probability Judgments for Fixed Values
Asa Palley, Duke University, The Fuqua School of Business, 100
Fuqua Drive, Box 90120, Durham, NC, 27708, United States of
America, asa.palley@duke.edu, Saurabh Bansal
■ TA64
64-Room 113A, CC
The Journey to Organizational Decision Quality (ODQ)
In many managerial decision problems, the distribution for a continuous random
variable must be obtained from expert judgments. Using a scale-free model of
judgmental errors, we present a method for estimating distribution parameters
through linear combinations of the judgments provided, where the weights are
explicit functions of the expert’s errors. Finally, we demonstrate the application
and benefits of our approach using data collected in an experimental study.
Sponsored Session
Chair: Carl Spetzler, CEO, Strategic Decisions Group, 745 Emerson
Street, Palo Alto, CA, 94301, United States of America,
cspetzler@sdg.com
1 - Progress in the Adoption of ODQ (Organizational Decision Quality)
Carl Spetzler, CEO, Strategic Decisions Group, 745 Emerson
Street, Palo Alto, CA, 94301, United States of America,
cspetzler@sdg.com
4 - A Turning Point Model Based on Exponential Smoothing
Xiaojia Guo, University College London, Dept. of Managment and
Innovation, Gower Street, London, WC1E 6BT, United Kingdom,
x.guo.11@ucl.ac.uk, Casey Lichtendahl, Yael Grushka-Cockayne
We propose a turning point model that extends the damped multiplicative trend
exponential smoothing model. Our model offers the ability to dynamically update
the local level and the growth trend, and ultimately to predict the turning point.
This dynamic turning point model can be contrasted with non-dynamic models
that are popular in the literature, such as the Bass diffusion model. We fit the
model to several well-studied time series and examine the model’s performance.
To set the stage for the following speakers, this session will provide a quick review
of ODQ, the ODQ maturity curve, and the progress that companies are making on
the journey to ODQ. After the following speakers, we will have a panel discussion
on the challenges faced by champions on the journey to ODQ and how they are
best met.
279
TA66
■ TA66
2 - Exact Formulations and Algorithms for the Pollution
Routing Problem
Yongjia Song, Virginia Commonwealth University,
821 W Franklin Street, Richmond, VA, United States of America,
ysong3@vcu.edu, Ricardo Fukasawa, Qie He
66-Room 113C, CC
Airline/Airport Operations Management
Sponsor: Aviation Applications
Sponsored Session
We propose for the first time exact formulations of the pollution routing problem.
These formulations are all mixed integer convex programs, with one being a
mixed integer second-order cone program. The lower bounds provided by the
continuous relaxations of these formulations are compared theoretically. Based
on our formulations, instances with up to 25 customers in the literature are
solved to optimality for the first time.
Chair: Ahmed Ghoniem, Isenberg School of Management, UMass
Amherst, 121 Presidents Dr., Amherst, MA, 01002, United States of
America, aghoniem@isenberg.umass.edu
1 - A Simulation-optimization Approach for Robust Aircraft Routing
and Flight Retiming
Mohamed Haouari, Professor, Qatar University, BP 2713, Doha,
Qatar, mohamed.haouari@qu.edu.qa, Mohamed Ben Ahmed,
Farah Zeghal Mansour
3 - A Column Generation Algorithm to Solve the Pollution
Routing Problem
Fernando Santos, PhD, University of Waterloo, 200 University
Avenue West, Waterloo, Canada, fernandoafonso1@gmail.com,
Qie He, Ricardo Fukasawa, Yongjia Song
We propose a novel simulation-optimization approach for solving the robust
aircraft routing and flight retiming problem. The approach requires iteratively
solving a mixed-integer quadratic programming problem that aims at optimally
inserting buffer times between consecutive flights, and invoking a Monte-Carlo
procedure for assessing the robustness of the generated schedules. We present the
results of extensive computational experiments that were carried out on a real
data.
We introduced a set partitioning formulation and a column generation algorithm
to solve the Pollution Routing Problem (PRP). To price out negative reduced cost
routes we proposed a labelling algorithm that derives novel dominance rules in
order to prune out unpromising labels and perform faster.
4 - The Deterministic Dispatch Waves Problem
Mathias Klapp, PhD Student, Georgia Tech, 755 Ferst Drive NW,
Main Building #326, Atlanta, GA, 30332-0205, United States of
America, maklapp@gatech.edu, Alan Erera, Alejandro Toriello
2 - Airlines’ Hedging Policies: An Empirical Approach to the U.S.
Domestic Market
Soheil Sibdari, Associate Professor, UMass Dartmouth,
285 Old Westport Rd, North Dartmouth, MA, Dartmouth,
United States of America, ssibdari@umassd.edu
We study last-mile delivery systems by formulating the deterministic dispatch
waves problem (DWP) that models a distribution center where geographically
positioned orders arrive at known action periods (waves) throughout the day. At
each wave, the decision maker chooses whether to dispatch a single vehicle or
not and the subset of open orders to serve in the vehicle’s route, with the
objective of minimizing operational costs and penalties for unserved requests.
We study airlines’s hedging policies during years 2002-2015 according to their
corporate’s yearly report. An empirical study examines airlines’ policy and
determine the impact of airline size, market share, and the airlines’ aircraft sizes
on the hedging effectiveness.
3 - Meta-heuristic Algorithm for the Multiple Runway Aircraft
Scheduling Problem
Bulent Soykan, Old Dominion University, Dep. of Eng. Mngt.
and Systems Eng., Norfolk, VA, United States of America,
bsoyk001@odu.edu, Ghaith Rabadi
■ TA68
68-Room 201B, CC
Joint Session TSL/Public Sector: Resilience in
Interdependent Infrastructure System
Multiple Runway Aircraft Scheduling Problem involves assigning both landing
and taking-off aircrafts to runways, sequencing them on each runway and
assigning each aircraft a landing or take-off time while considering predetermined
time windows for each aircraft to land or take-off. This research aims to develop a
tabu search/path relinking algorithm for the static case of the problem, where all
information of aircraft is known in advance.
Sponsor: Transportation, Science and Logistics
Sponsored Session
Chair: Mohammad Khodayar, Southern Methodist University, 6251
Airline Rd, Junkins Bldg, suite 334, Dallas, TX, 75275, United States of
America, mkhodayar@mail.smu.edu
1 - Interdiction Analysis of Coupled Electricity and Natural
Gas Networks
Bowen Hua, The University of Texas at Austin,
1616 Guadalupe St, Austin, TX, 78701, United States of America,
bhua@utexas.edu, Ross Baldick
4 - A Two-Stage Airport Surface 4D Taxiing Trajectory Scheduling
Strategy Considering Runway Exit Select
Xiang Zou, Tsinghua University, Room 430, Main Building,
Tsinghua Univ., Beijing, China, x-zou10@mails.tsinghua.edu.cn,
Bang An
This paper proposes a two-stage airport taxing scheduling policy. In the first stage,
all of the interested aircrafts are assigned initial routes. Then, aircrafts unavailable
to fulfill their initially assigned routes are rescheduled. We do not fix the runway
exits of landing aircrafts. Instead, we introduce Runway Exit Availability and a
MIP model to assign 4D taxiing trajectories. Test in the environment of Beijing
Capital Airport shows the effectivity and efficiency of the approach.
We present a bilevel optimization model to identify the critical components of the
coupled power and natural gas pipeline system. The upper-level problem involves
the interdiction decisions and the lower-level problem represents the operation of
the coupled system. We model the system operation as an MILP to include the
nonlinear flow-pressure relations. Some theoretical properties of this bilevel
program are analyzed. A decomposition algorithm is proposed to solve the
problem.
■ TA67
2 - Quantifying the Resilience of an Urban Traffic – Electric Power
Coupled System
Elise Miller-Hooks, Professor, University of Maryland, College
Park, MD, elisemh@umd.edu, Seksun Moryadee, Steven Gabriel,
Hossein Fotouhi
67-Room 201A, CC
Advanced Routing Models
Sponsor: TSL/Freight Transportation & Logistics
Sponsored Session
A nonlinear, stochastic, mixed integer program is presented for quantifying the
resilience of the coupled traffic-power network to a disruption. The model
captures interdependencies in this system, and seeks an optimal allocation of
limited mitigation, preparedness and response resources to obtain an efficient
resource allocation plan and maximum resilience estimate.
Chair: Qie He, University of Minnesota, 111 Church Street SE,
Minneapolis, United States of America, qhe@umn.edu
1 - Pollution-routing Problems with Speed and Departure
Time Optimization
Raphael Kramer, PhD Student, Universita degli Studi di Modena
e Reggio Emilia, Via Amendola 2, 42122, Reggio Emilia, Italy,
raphael.kramer@unimore.it, Thibaut Vidal, Anand Subramanian,
Nelson Maculan
3 - Improving the Resilience of Multiple Energy Carrier Microgrids
Against Deliberate Disruptions
Saeed Dehghan Manshadi, Southern Methodist University,
6251 Airline Rd, Junkins Bldg, Dallas, TX, 75275,
United States of America, manshadi@mail.smu.edu,
Mohammad Khodayar
We consider the Pollution-Routing Problem with possible departure time
optimization. This enables to better allocate human resources to time periods with
higher delivery needs. An algorithm for speed and departure time optimization is
introduced for any fixed route. Its optimality is proven. Integrating this algorithm
into a classical metaheuristic generates high-quality routing solutions.
Experimental analyses show the impact of departure time on speeds decision,
emissions and labor costs.
This paper proposes a framework to identify the vulnerable components in the
coordinated natural gas and electricity distribution networks in microgrids and to
ensure the resilient operation of such interdependent networks. The proposed
framework addresses deliberate actions to disrupt the energy flow in the
microgrids and proposes reinforcement strategies to increase the resilience of the
energy supply.
280
■ TA69
TA72
2 - Failure Prediction and Sensor Spacing Optimization Along
Track Corridors
Yanfeng Ouyang, Univ. Of Illinois, 205 N. Mathews Ave, Urbana,
United States of America, yfouyang@illinois.edu, Zhaodong Wang
69-Room 201C, CC
Maritime Logistics
This talk describes a machine-learning based framework for determining sensor
deployment to ensure optimal reporting of potential incident-prone failures of the
passing traffic. A simulation-based optimization model is used to find the optimal
sensor spacing.
Sponsored Session
Chair: Irina Benedyk, United States of America, birina@purdue.edu
1 - Solving the Pre-Marshalling Problem to Optimality
Kevin Tierney, Assistant Professor, University of Paderborn,
Warburger Strafle 100, Paderborn, 33098, Germany,
kevin.tierney@upb.de, Stefan Voss, Dario Pacino
3 - Development and Application of Line-of-road Emulator Tool
in CSX
Yu Wang, Manager Operations Research, CSX Transportation
Inc., 500 Water Street, Jacksonville, FL, 32202,
United States of America, Yu_Wang@csx.com, Eric Pachman
The pre-marshalling problem is a key problem at container terminals. The goal is
to find a minimal sequence of re-shuffling containers in a set of stacks such that
they are arranged according to the time each container must leave the stacks. We
present a novel algorithm using A* and IDA* combined with several novel
branching and symmetry breaking rules. We solve over 500 previously unsolved
benchmark instances to optimality clearly outperforming current state-of-the-art
methods.
Line-of-Road Emulator is a web-based tool to visualize train movements in a GIS
view. The tool can highlight slow-moving and/or long-dwell trains with different
styles of bubbles, which provides informative insights to help railroad managers
understand the situation and investigate the reasons causing congestions. The tool
was used to create an illustration video about the congestion happened on the
northern tier of CSX network in 2014 winter, and has received high evaluation
from the users.
2 - A Genetic Algorithms Based Approach to Develop Cost-Effective
Annual LNG Delivery Program
Fatih Mutlu, Asst. Professor, Qatar University, Doha, Doha, Qatar,
fatihmutlu@qu.edu.qa
4 - Optimization Algorithms for Hump Yard Decision Support System
Alexey Sorokin, Senior Systems Engineer, Optym, 7600 NW 5th
Place, Gainesville, FL, 32607, United States of America,
alexey.sorokin@optym.com, Ravindra Ahuja, Krishna Jha
Developing a cost effective annual delivery program for liquefied natural gas
suppliers is known to be among the most challenging integrated inventory,
production, and maritime delivery routing problems. We use a genetic algorithms
based approach to solve this problem. We produce alternative routes for the
vessels, each of which represents a chromosome. Our method performs better
than the exact solution method in all of the problem instances we solved.
Rail cars are classified to their appropriate outbound trains in yards. Important
decisions made by yardmasters include the order in which trains should be
humped and classification track on which a block should be built at any point in
time. We developed optimization modules for a real-time decision support system
that can assist yardmasters with these decision. Benefits of the optimization
algorithms were computed using a hump-yard simulation system previously
developed by Optym.
3 - A Mathematical Model for the Ship Scheduling and Cargo
Assignment Problem
Salomon Wollenstein Betech, Student, Instituto Tecnologico de
Estudios Superiores de Monterrey, Av Carlos Lazo 100, Alvaro
Obregón, DF, 01389, Mexico, s.wollenstein@gmail.com
■ TA72
72-Room 203A, CC
Middle-size companies with maritime shipping face a scheduling and cargoassignment problem. Given a set of demands, suppliers, contracts, and ships, the
company must design their operations to minimize cost. A mathematical model is
proposed that simultaneously solves the ship scheduling and cargo assignment
problem for a period of a year, discretizing time in days. The algorithm is capable
of solving the problem at a rate of five ships and ports in ten minutes.
DDDAS for Industrial and System Engineering
Applications I
Sponsor: Quality, Statistics and Reliability
Sponsored Session
4 - A Bivariate Probit Model to Analyze Perspectives for Container
Shipping on the Northern Sea Route
Irina Benedyk, United States of America, birina@purdue.edu,
Srinivas Peeta
Chair: Shiyu Zhou, Professor, University of Wisconsin-Madison,
Department of Industrial and Systems Eng, 1513 University Avenue,
Madison, WI, 53706, United States of America, shiyuzhou@wisc.edu
This study seeks to explore opportunities and barriers for container freight
shippers to use the Northern Sea Route. A stated preference survey of freight
shippers in East Asia and Europe is conducted. A Bivariate Probit Model is used to
investigate attitudes towards the usage of the North Sea Route, and identify key
factors that influence them.
Co-Chair: Yu Ding, Professor, Texas A&M University, ETB 4016,
MS 3131, College Station, TX, United States of America,
yuding@iemail.tamu.edu
1 - Dynamic Data Driven Applications Systems DDDAS):
New Capabilities in Data Analytics
Frederica Darema, Program Director, Air Force Office of Scientific
Research, United States of America, frederica.darema@us.af.mil
■ TA70
This talk provides an overview of future directions enabling in new
methodologies for analytics through the DDDAS (Dynamic Data Driven
Applications Systems) paradigm. We will discuss how DDDAS allows new
capabilities in data analytics to enable optimized and fault tolerant systems
management, improved analysis and prediction of system conditions, in a diverse
set of application areas ranging from aerospace applications to smart cities, to
manufacturing planning and control, and cybersecurity.
70-Room 202A, CC
Advanced Analytics in Tactical Decision Making
Sponsor: Railway Applications
Sponsored Session
Chair: Krishna Jha, Vice President Research And Development, Optym,
7600 NW 5th Place, Gainesville, FL, 32607, United States of America,
krishna.jha@optym.com
1 - Forecast Locomotive Surplus and Deficit to Balance the
Terminals and Shops
Kamalesh Somani, CSX Transportation, 500 Water St,
Jacksonville, FL, 32202, United States of America,
Kamalesh_Somani@CSX.com, Shankara Kuppa,
Artyom Nahapetyan
2 - Offline Learning for Dynamic Data-driven Capability Estimation for
Self-aware Aerospace Vehicles
Douglas Allaire, Assistant Professor, Texas A&M University,
425 MEOB, 3123 TAMU, College Station, TX, 77843,
United States of America, dallaire@tamu.edu, Benson Isaac
A self-aware aerospace vehicle can dynamically adapt the way it performs
missions by gathering information about itself and its surroundings and
responding intelligently. We present an information-theoretic approach to offline
learning via the optimization of libraries of strain, capability, and maneuver
loading using physics-based computational models. Online capability estimation is
then achieved using by a Bayesian classification process that fuses dynamic,
sensed data.
Number of locomotives coming into a terminal may not be same as number of
locomotives going out. This creates imbalance where some terminals are in
constant need for locomotives and some other terminals usually have spare
locomotives. Similarly a shop may receive more locomotive than its capacity and
at the same time another shop may not be used to its full capacity. We developed
advance analytics tools which help to minimize network balancing cost and any
train delay because of locomotives.
281
TA73
■ TA73
2 - Concurrent Process Plan Optimization for Additive Manufacturing
Bahir Khoda, Professor, North Dakota State University, Room #
202F Civil and Industrial Enginee, 1410 14th Avenue North,
Fargo, ND, 58102, United States of America,
akm.khoda@ndsu.edu, Amm Nazmul Ahsan, Md Habib
73-Room 203B, CC
Functional Data Analysis
Sponsored Session
Implementing additive manufacturing processes effectively requires addressing
issues of process proficiencies and resource utilization, both of which have a
strong environmental impact. In this paper, both part build orientation and
material deposition direction are concurrently optimized by analyzing part
geometry to minimize the resource requirement. A concurrent multicriteria
process plan optimization framework is developed using Genetic Algorithms (GA)
technique.
Chair: Moein Saleh, Discover Financial Services/ Arizona State
University, 699 S Mill Ave, Tempe, AZ, 85281, United States of
America, Moein.Saleh@asu.edu
1 - On the Use of Gaussian Processes for Surface and Profile Data
Enrique Del Castillo, Penn State University, Industrial Eng. and
Statistics Depts., State College, United States of America,
exd13@psu.edu
3 - Online Sensor-based Monitoring in Aerosol Jet Printing Process
Prahalad Rao, SUNY Binghamton, 4400 Vestal Pkwy. E,
Binghamton, NY, United States of America,
prao@binghamton.edu, Roozbeh Salary, Jack Lombardi,
Mark Poliks
Standard applications of Gaussian Processes in manufacturing data have
traditionally been based on models of the form z(x,y) where x,y,z are coordinates
acquired with some sensor, so correlation is assumed to occur on euclidean space
external to the surface. We show new methodology that assumes instead
correlation exists on the intrinsic surface points along geodesic distances, and
show how this leads to better surface reconstruction in both simulated and real
datasets.
Aerosol Jet Printing (AJP) is an additive manufacturing process (AM) is emerging
as a viable method for printing conformal electronics. However, teething quality
related problems in AJP remain unresolved. We propose approaches based on
image processing and sensor data analytics to achieve online quality monitoring
in the AJP process. The effectiveness of the proposed approach is assessed and
evaluated with several real case studies implemented on an aerosol jet printer
setup.
2 - Functional Clustering with Applications in Single
Molecule Experiments
Ying Hung, yhung@stat.rutgers.edu
Cell adhesion experiments refer to biomechanical experiments that study protein,
DNA, and RNA at the level of single molecules. Motivated by analyzing a single
molecule experiment, a new statistical framework is proposed based on functional
clustering approaches. Simulations and applications to real experiments are
conducted to demonstrate the performance of the proposed method.
■ TA75
75-Room 204B, CC
IBM Research Best Student Paper Award I
3 - Design of Experiments for Functional Response
Moein Saleh, Discover Financial Services/ Arizona State
University, 699 S Mill Ave, Tempe, AZ, 85281,
United States of America, Moein.Saleh@asu.edu, Rong Pan
Sponsor: Service Science
Sponsored Session
Chair: Ming-Hui Huang, National Taiwan University, Taiwan - ROC,
huangmh@ntu.edu.tw
1 - Best Student Paper Competitive Presentation
Ming-Hui Huang, National Taiwan University, Taiwan - ROC,
huangmh@ntu.edu.tw
Applications of DOE for single response variable can be seen in nearly every
disciplines in science and engineering. However, there are very few publications
that discussed optimal design for the experiments with multiple responses taken
over different points of a continuum variable. This continuum can be any other
continuous variable for functional data analysis such as time in longitudinal
study. My study focuses on developing a framework for designing the
experiments for functional response.
Finalists of the IBM Research Best Student Paper Award present their research
findings in front of a panel of judges. The judging panel will decide the order of
winners, which will be announced during the business meeting of the Service
Science Section at the Annual Conference.
4 - Monitoring and Diagnostics of High Dimensional
Multi-stream Data
Samaneh Ebrahimi, Research Assistant, Georgia Institute of
Technology, 755 Ferst Drive, Atlanta, GA, 30332, United States of
America, samaneh.ebrahimi@gatech.edu, Kamran Paynabar,
Chitta Ranjan
1- Service Innovation and the Role of Collaboration
Cong Feng, Syracuse University, 721 University Avenue,
Syracuse NY, United States of America, feng@congfeng.net,
K. Sivakumar
Results show that (1) the effect of service innovation on firm performance is
greater for service firms than manufacturing firms; (2) the relationship between
the propensity for service innovation and three types of collaboration is significant; and (3) vertical and third-party collaborations are more beneficial than horizontal collaboration for service firms.
Correlated high-dimensional data streams (HDDS) pose significant challenges in
Statistical Process Monitoring. In this research, we integrate PCA and Adaptive
Lasso, and propose a novel approach for effective process monitoring and
diagnosis of HDDS. The effectiveness of the proposed approach is validated
through simulation and a case study.
2 - Brand Equity and Extended Service Contract Purchase Decisions
Moein Khanlari Larimi,University of Alberta, Canada,
khanlari@ualberta.ca, Paul Messinger
■ TA74
In this paper, we explore the role of brand equity on consumers’ extended service contract (ESC) purchase decisions. We draw from past findings to show that
higher brand equity has an overall positive impact on ESC purchase decisions.
We also explore the positive impact of stores on ESC purchase decisions.
74-Room 204A, CC
System and Process Informatics in Additive
Manufacturing (I)
3 - Regulating Greed over Time
Stefano Traca, Massachusetts Institute of Technology, Cambridge,
MA, United States of America, stet@mit.edu, Cynthia Rudin
Sponsored Session
In retail, there are predictable yet dramatic time-dependent patterns in customer
behavior, such as periodic changes in the number of visitors, or increases in visitors just before major holidays (e.g., Christmas). The current paradigm of multiarmed bandit analysis does not take these known patterns into account, which
means that despite the firm theoretical foundation of these methods, they are
fundamentally flawed when it comes to real applications. This work provides a
remedy that takes the time-dependent patterns into account, and we show how
this remedy is implemented in the UCB and e-greedy methods. In the corrected
methods, exploitation (greed) is regulated over time, so that more exploitation
occurs during higher reward periods, and more exploration occurs in periods of
low reward. In order to understand why regret is reduced with the corrected
methods, we present a set of bounds that provide insight into why we would
want to exploit during periods of high reward, and discuss the impact on regret.
Our proposed methods have excellent performance in experiments, and were
inspired by a high-scoring entry in the Exploration and Exploitation 3 contest
using data from Yahoo! Front Page. That entry heavily used time-series methods
to regulate greed over time, which was substantially more effective than other
contextual bandit methods.
Chair: Linkan Bian, Assistant Professor, Mississippi State University, 260
McCain Building, Mississippi State, Starkville, MS, 39762, United States
of America, bian@ise.msstate.edu
1 - Accelerated Process Optimization for Laser-based Additive
Manufacturing (LBAM)
Amir M. Aboutaleb, Mississippi State University, 260 McCain
Building, Mississippi State, MI, 39762, United States of America,
aa1869@msstate.edu, Linkan Bian, Alaa Elwany, Nima Shamsaei,
Scott M. Thompson
A novel Design-of-Experiment methodology is proposed to efficiently optimize
process control parameters for LBAM by leveraging data obtained from prior
related but non-identical studies. Our method accounts for unavoidable difference
between the experimental conditions of the current and prior studies and
quantify the associated uncertainty, which is further updated using real-world
data generated in the current study.
282
TA78
■ TA77
4 - Assessing the Impact of Product and Service Quality on
Consumer Returns: A Data Analytics Study
Necati Ertekin,Texas A&M University, Mays Business School,
College Station TX 77840, United States of America,
nertekin@mays.tamu.edu, Gregory Heim, Michale Ketzenberg
77-Room 300, CC
Green Supply Chain Management
Contributed Session
We contribute to the understanding of consumer return behavior by examining
the association between in-store customer experience during a purchase and a
subsequent return. We demonstrate that retail efforts such as increasing salesperson competence and improving store environment that are so long believed to
prevent returns may indeed induce returns.
Chair: Vinay Gonela, Assistant Professor Of Management, Southwest
Minnesota State University, CH 214, 1501 State Street, Marshall, MN,
56258, United States of America, vinay.gonela@smsu.edu
1 - The Impact of Contracts on Environmental Innovation in a
Supply Chain
Seyoun Jung, PhD Student, KAIST (Korea Advanced Institute of
Science and Technology), 85 Hoegiro, Dongdaemun-gu, Seoul,
Korea, Republic of, ssebea@business.kaist.ac.kr, Bosung Kim,
Kun Soo Park
■ TA76
76-Room 204C, CC
Advances in Simulation-based Optimization I
We examine the impact of contracts between a supplier and a manufacturer on
the supplier’s environmental innovation. We calculate and compare the
equilibrium outcomes under three types of contract such as wholesale-price,
revenue-sharing, and quality-dependent contracts.
Sponsor: Simulation
Sponsored Session
Chair: Jie Xu, George Mason University, 4400 University Dr., MS 4A6,
Engr Bldg, Rm 2100, Fairfax, VA, 22030, United States of America,
jxu13@gmu.edu
1 - Estimating the Probability of Convexity of a Function Observed
with Noise
Nanjing Jian, PhD Student, Operations Research and Information
Engineering, 288 Rhodes Hall, Cornell University, Ithaca, NY,
14850, United States of America, nj227@cornell.edu,
Shane Henderson
2 - Producer-dominated Green Supply Chain Collaboration under
Trade-in Programs
Chih-Tien Chiu, Doctoral Student, National Taiwan University,
No.1,Sec. 4, Roosevelt Rd., Taipei, 10617, Taiwan - ROC,
d03741001@ntu.edu.tw, Mu-chen Chen, Jiuh-biing Sheu
This paper aims to address new-product/used-product pricing in a green logistics.
We adopt the dynamic programming approach integrated with the logit model to
formulate the n-period trade-in pricing-logistics problem, where the logit model is
utilized for trade-in service channels choice. Data collected via stated preference
experiments are used for the parameter estimation of the logit model, followed by
conducting quantitative analyses to provide important findings and managerial
insights.
Given estimates of the values of a function observed with noise from simulation
on a finite set of points, we wish to sequentially estimate the probability that the
function is convex. By updating a Bayesian posterior on the function values, we
iteratively estimate the posterior probability of convexity by solving certain linear
programs in a Monte Carlo simulation. We discuss a variety of variance reduction
methods for the estimation and the linear programs associated with each.
3 - Metrics for Sustainable Operations: Current State and
Path to Improvement
Remi Charpin, Clemson University, 100 Sirrine Hall, Clemson,
United States of America, rcharpi@g.clemson.edu, Aleda Roth
2 - Adaptive Sampling Trust Region Optimization
Sara Shashaani, Associate Professor, Department of Statistics,
Purdue University, 250 N University Street, West Lafayette, IN,
47907, United States of America, pasupath@purdue.edu,
Raghu Pasupathy
From an operations and supply chain management lens, we examine
sustainability metrics currently being reported by firms. We propose that certain
metrics are ‘attractors,’ as they are apt to lead the business towards sustainability,
whereas others are deemed to be ‘detractors’ that are likely to be used for
‘greenwashing.’
We develop derivative free algorithms for optimization contexts where the
objective function is observable only through a stochastic simulation. The
algorithms we develop follow the trust-region framework where a local model is
constructed, used, and updated as the iterates evolve through the search space.
We incorporate adaptive sampling to keep the variance and the squared bias of
the local model in lock step, in a bid to ensure optimal convergence rates.
4 - Stochastic Optimization of Sustainable Industrial Symbiosis
Based Hybrid Generation Bioethanol Supply Chain
Vinay Gonela, Assistant Professor Of Management, Southwest
Minnesota State University, CH 214, 1501 State Street, Marshall,
MN, 56258, United States of America, vinay.gonela@smsu.edu,
Atif Osmani, Jun Zhang, Joseph Szmerekovsky
3 - Parallel Empirical Stochastic Branch & Bound
Sajjad Taghiye, George Mason University, 4400 University Dr.,
MS 4A6, Engr Bldg, Rm 2100, Fairfax, VA, 22030,
United States of America, staghiy2@gmu.edu, Jie Xu
This paper focuses on designing a new industrial symbiosis based hybrid
generation bioethanol supply chain (ISHGBSC). A SMILP model is proposed to
design the optimal ISHGBSC under different sustainability standards. The result
provides guidelines for policy makers to determine the appropriate standard to
use under different sustainable concerns. In addition, it provides investors a
guideline to invest in different technologies under different sustainability
standards.
To efficiently solve problems with time-consuming high-fidelity simulations, we
develop a new parallel algorithm known as parallel empirical stochastic branch &
bound (PESBB) to exploit the power of high performance computing. We will
discuss synchronous and asynchronous versions of PESBB and present initial
numerical results to demonstrate the scalability of PESBB.
4 - Finding the Best using Multivariate Brownian Motion
Seong-hee Kim, Professor, Georgia Institute of Technology,
755 Ferst Dr NW, Atlanta, GA, 30332, United States of America,
skim@isye.gatech.edu, Ton Dieker, Seunghan Lee
■ TA78
78-Room 301, CC
Big Data and Energy
We present a new fully sequential procedure based on multivariate Brownian
motion when variances are known but unequal. The procedure uses an ellipsoid
as a continuation region, and a system with the worst sample mean is eliminated
whenever the procedure’s statistic exits the ellipsoid. The size of the ellipsoid
changes as the number of survivors decreases. Experimental results are provided
for both equal and unequal variances.
Contributed Session
Chair: Feng Gao, SGRI North America, 5451 Great America Parkway,
Santa Clara, CA, 95054, United States of America, feng.gao@sgrina.com
1 - Resilient Power System State Estimation using
Compressive Sensing
Hanif Livani, Assistant Professor, University of Nevada Reno,
Electrical & Computer Engineering, MC 0111, 1185 Perry St, /
Room 302, Reno, NV, 89557, United States of America,
hlivani@unr.edu
Phasor Measurement Units (PMU) have become widely used for power system
monitoring and control. However, they are not installed on all the buses in a
network. Therefore, PMU-only state estimation encounters problems arising from
a limited number of installed PMUs and probable data losses as the results of
congestion or disconnection in communications.In this study, we propose power
system state estimation using Compressive Sensing (CS) algorithm which is
resilient to loss of data.
283
TA79
Tuesday, 11:00am - 12:30pm
2 - Data Exploration of Publicly Reported Power Outages to Assess
Grid Reliability and Damages
Michael Sohn, Staff Scientist, Lawrence Berkeley National
Laboratory, One Cyclotron Road, Mail Stop: 90R2008, Berkeley,
CA, 94720, United States of America, mdsohn@lbl.gov,
Joseph Eto, Kristina Lacommare, Laurel Dunn
■ TB01
We have amassed a database of power outages, with high temporal resolution,
from across the US. We present an analysis of the database, focusing on the
distribution of outages by duration, customers affected, location, time, etc. We
also link the data to orthogonal datasets to estimate the types of customers
affected for a particular outage. Finally, we estimate the cost of power
interruptions, and discuss implications on the reliability costs nationwide.
Cyber and Logistics Applications
Sponsored Session
Chair: Natalie Scala, Assistant Professor, Towson University, Dept. of eBusiness and Tech Management, 8000 York Road, Towson, MD, 21252,
United States of America, nscala@towson.edu
1 - Operations Research Initiatives in Cyber Defense
Paul Goethals, Army Cyber Institute, United States Military
Academy, West Point, NY, United States of America,
paul.goethals@usma.edu
3 - Energy Disaggregation Based on Stochastic Dynamic
Programming with Collocation Method
Feng Gao, SGRI North America, 5451 Great America Parkway,
Santa Clara, CA, 95054, United States of America,
feng.gao@sgrina.com, Chris Saunders, Yang Yu, Wendong Zhu,
Guangyi Liu
The purpose of energy disaggregation is to separate energy consumption for a
consumer into the energy data for individual appliances. The benefit lies in the
promotion of improved consumption behaviors and adaptation of energy-efficient
devices. The paper presents a dynamic model for devices power consumption;
considers uncertainty of consumption; and proposes a fix-point iterative schema
to efficiently resolve the stochastic problem. The paper demonstrates its result on
a simulated data set.
Despite the reduction in the total Army population, the cyber force structure
continues to grow in strength and impact. This presentation describes a number
of Operations Research initiatives that could benefit the cyber community.
Research trends and areas of future work are also offered.
2 - Automated Identification Technology Devices for Naval Seabasing
Natalie Scala, Assistant Professor, Towson University, Dept. of eBusiness and Tech Management, 8000 York Road, Towson, MD,
21252, United States of America, nscala@towson.edu, Jennifer
Pazour
4 - Transmission Planning with Renewable Distributed
Generation Uncertainty
Fikri Kucuksayacigil, Iowa State Uni. Industrial Engineering,
3004 Black Engineering, Ames, IA, 50011, United States of
America, fksayaci@iastate.edu, Kyung Jo Min
We present a value focused decision model for naval seabasing. We discuss
automated identification technology devices as alternatives to a multi-objective
decision model with the goal of selecting the preferred device for seabasing
logistics support. Criteria for this model include metrics and associated measures
related to seabasing.
There have been substantial developments of distributed generation from
renewable energy sources. This has created new challenges in transmission
planning as distributed generation leads to uncertainties on the use of
transmission lines. To address this uncertainties, we utilize a jump-diffusion
demand process and binomial lattice to show how the best transmission is
planned under the risk of self-supporting communities. From the resulting
analysis, economic implication and managerial insights will be discussed.
3 - Logistics Engineering Solution for Reverse-engineering Topology
Alan Briggs, INFORMS Maryland, 8606 Aspen Grove Court,
Odenton, MD, 21113, United States of America,
awbriggs@gmail.com
Using monte carlo simulation, author uses proximate location data to reverse
engineer network topology.
■ TA79
79-Room 302, CC
■ TB02
Software Demonstration
Cluster: Software Demonstrations
Invited Session
1 - SigmaXL, Inc. - What’s New in SigmaXL® Version 7
John Noguera, CTO & Co-founder, SigmaXL, Inc.
Invited Session
Homeland Security Decision Making
SigmaXL is a user friendly Excel Add-In tool for Process Improvement, Six Sigma
Quality and Statistics. We introduce SigmaXL and the new features in Version 7:
“Traffic Light” Automatic Assumptions Check for T-tests and ANOVA, Automatic
Normality Check for Pearson Correlation and Small Sample Exact Statistics for
One-Way Chi-Square, Two-Way (Contingency) Table and Nonparametric Tests.
Exact statistics are appropriate when the sample size is too small for a Chi-Square
or Normal approximation to be valid.
Chair: Jun Zhuang, University at Buffalo, SUNY, 317 Bell Hall, Buffalo,
NY, 14221, United States of America, jzhuang@buffalo.edu
Co-Chair: Fei He, Assistant Professor, Texas A&M University-Kingsville,
700 University Blvd., Kingsville, TX, 78363, United States of America,
fei.he@tamuk.edu
1 - Multi-Objective Optimization Models in Urban Security
Jose Emmanuel Ramirez-Marquez, Associate Professor, Stevens
Institute of Technology, 1 Castle Point Rd, Hoboken, NJ, 07030,
United States of America, jmarquez@stevens.edu,
Mohammed Muaafa
2 - Mathworks - MATLAB: An Environment for Operations Research
and Data Analytics
Seth DeLand, MathWorks, Data Analytics. Product Manager
MATLAB is a platform for analysis, visualization, simulation, and optimization.
You can access and analyze real-world data and develop customized algorithms
that scale to your largest problems. Join us to see how MATLAB can help you
explore data, develop algorithms, and integrate analytics into enterprise
applications. You’ll also learn about new features including mixed-integer linear
programming, machine learning, and working with Big Data.
The allocation of limited resources is a daily dilemma for police commanders.
More than 50% of police department costs go into patrolling operations, which
include responding to emergencies and maintaining police presence. This study
aims to address the challenging tradeoff police face when designing patrolling
strategies between lessening the economic burden of crime prevention and
maintaining high levels of public safety.
2 - A Literature Review of Recent Attacker-defender Games
Fatemeh Mousapour, SUNY Buffalo, 338 Bell Hall, Buffalo, NY,
United States of America, f.moosapoor@yahoo.com, Jun Zhuang
This research provides an extensive review of game-theoretic analysis of attackerdefender models. Those models are categorized according to different defense
measures, attack tactics, system structures, game types, player rationality and risk
preferences. Statistical charts and tables are presented to identify patterns and
trends in this area. Through the content analysis framework, some research gaps
and future research directions are identified.
284
3 - Protecting and Restoring Facilities from Intentional Attacks
Chi Zhang, Assistant Professor, Tsinghua University, 100084
Beijing, Department of Industrial Engineering, Beijiing, China,
czhang@tsinghua.edu.cn, Sachuer Bao
TB05
4 - The Value of Flexibility and Shift Extensions in
Physician Scheduling
Andreas Fögener, University of Augsburg, Universitätsstrafle 16,
Universität Augsburg, WiWi, Augsburg, De, 86159, Germany,
andreas.fuegener@unikat.uni-augsburg.de, Jens Brunner
Besides protecting facilities from intentional attacks, another paramount issue is
taken into consideration- restoring destroyed facilities to optimal service level
within a given time interval. The defender decides which facilities to protect
before an attack, and resource allocation between improving capacities of
operational facilities and rebuilding destroyed facilities after an attack, to
maximize profit by satisfying customer demands. The problem is solved by an ant
colony algorithm.
Scheduling physicians is a relevant topic in hospitals. In the literature, demand is
usually assumed to be deterministic. However, surgery durations and emergencies
contain uncertainty. We model stochastic physician demand using a scenariobased approach. We introduce flexible shift extensions, where physicians might
have to work longer to match supply with demand and simultaneously increase
predictability of working hours. We propose a mixed-integer model and a column
generation heuristic to solve our problem.
4 - Coordinating Pre- and Post-disaster Resource Allocation at
Multiple Locations
Fei He, Assistant Professor, Texas A&M University-Kingsville, 700
University Blvd., Kingsville, TX, 78363, United States of America,
fei.he@tamuk.edu, Jun Zhuang
■ TB04
Resource allocation in the face of disaster aims to improve the efficiency and
effectiveness of disaster relief. In this research, disaster preparedness and relief at
multiple locations are modeled in a two-stage stochastic programming framework
with the objective of loss minimization. New insights of coordinating
preparedness and relief at multiple locations are provided.
The Business of Music and Emotion in Social Media
Invited Session
5 - Model Validation for a Public-private Partnerships Model in
Disaster Management
Vineet Madasseri Payyappall, PhD Student, University at Buffalo,
305 Winspear Avenue (Upper), Buffalo, NY, 14215, United States
of America, vineetma@buffalo.edu, Peiqiu Guan, Jun Zhuang
Chair: Chris Smith, TRAC-MTRY, 28 Lupin Lane, Carmel Valley, 93924,
United States of America, cmsmith1@nps.edu
1 - Philippine Language and Emotion During Typhoon
Haiyan/Yolanda
Amanda Andrei, Graduate Student, Georgetown University,
aa1436@georgetown.edu
This research designed and conducted an experiment to validate a public-private
partnerships model in disaster management. A two-staged experiment was
conducted in a computer simulated environment. The risk behaviors of the
subjects were evaluated in the first stage, and the second stage collected the
decision of the subjects, who played the role of the private sectors, under different
scenarios. The experiment shows that our model results are consistent with the
experimental results.
An investigation of language and emotion in tweets from the Philippines before
and after 2013 supertyphoon Haiyan/Yolanda using Linguistic Inquiry and Word
Count (LIWC), breakpoint analysis, and a computational clustering tool revealed
differences in topics and emotions depending on whether messages were
expressed in English or Filipino.
2 - Subscribe or Sell: Itunes vs. Google Play Music all Access
Hooman Hidaji, PhD Student, University of Alberta, #1604 8515
112 St. NW, Edmonton, Al, T6G1K7, Canada,
hooman.hidaji@ualberta.ca
■ TB03
Recently, subscription has become a popular method of user monetization in
online media business along with selling model. It is expected that firms utilize
both approaches to cover as much demand as possible. However, pricing strategy
of the firms is crucial in determining the demand for the two. In this study, using
an economic model with endogenous demand, we set to model how the firm
decides on the business model. Different user types and business modeldependent demand are considered.
New Topics in Scheduling
Cluster: Scheduling and Project Management
Invited Session
Chair: Rainer Kolisch, rainer.kolisch@wi.tum.de
1 - Coordinating Subcontractor Scheduling with Divisible Jobs
Behzad Hezarkhani, Assistant Professor, Nottingham University
Business School, Jubilee Campus, Nottingham, United Kingdom,
behzad.hezarkhani@nottingham.ac.uk, Wieslaw Kubiak
3 - Stock Market Prediction using Disparate Data Sources
Bin Weng, Auburn University, 425 Opelika Rd Apt. 224, Auburn,
Al, 36830, United States of America, bzw0018@auburn.edu,
Fadel Megahed
We study a decentralized scheduling problem with a single subcontractor and
several agents having divisible jobs. Under complete information, we design
pricing schemes that always make the agents’ decisions coincide with efficient
schedules. Under private information, we prove that the pivotal mechanism
makes truth-telling the only optimal choice of the agents when announcing their
processing times. We comment on the subcontractor’s revenue under complete
and private information.
Stock market prediction has attracted much attention from academia as well as
business. In recent years, social media is considered as a new source to affect
human’s behavior and decision-making. In this paper, we will develop a new way
to predict the movement of the stock market using disparate data source, social
media data and market data. In order to predict the stock price more accurately,
the model is developed using multivariable selection method and machine
learning statistic methods.
2 - Single Machine Scheduling via Decision Theory
J.J. Kanet, Department of MIS/OM/DSC, 300 College Park,
University of Dayton, Dayton, OH, 45419-2130,
United States of America, Kanet@udayaton.edu
■ TB05
We consider the following procedure for scheduling a single machine. At time t
the machine is free with a set N of n jobs ready to occupy it. Thus, we have n
choices for jobs to occupy the machine starting at time t with the remaining n-1
jobs completed later. Given that a job k is tentatively chosen to next occupy the
machine, we calculate its completion time and the expected value (E) of the
completion times of the remaining n-1 jobs. We do this for each of the n choices,
producing for each the set of completion times C = {Cj?j?N}. We then evaluate the
objective Z = f(C) choosing that job k?Z is minimum to next occupy the machine.
We provide an unbiased estimator of the set C and show that the procedure
provides optimum results when the objective Z is to minimize flow time or
maximum tardiness.
Social Media in Business
Invited Session
3 - Scheduling on a SIngle Machine under Time of Use Tariffs
Kan Fang, Tianjin University, No 92 Weijin Road, Nankai District,
Tianjin, 300072, China, zjumath@gmail.com, Nelson Uhan
Chair: Dokyun Lee, Carnegie Mellon University, Pittsburgh, PA,
United States of America, leedokyun@gmail.com
1 - Understanding the Impact of Discussions on Quality of
Crowdsourced Content – The Case of Wikipedia
Srikar Velichety, PhD Student, Eller College of Management,
University of Arizona, 1130 E Helen St, Tucson, AZ, 85719,
United States of America, srikarv@email.arizona.edu,
Jesse Bockstedt, Sudha Ram
We consider the problem of scheduling jobs on a single machine to minimize the
total electricity cost of processing these jobs under time-of-use electricity tariffs.
We show the computational complexity of this problem for both the uniformspeed and speed-scaling cases, present different approximation algorithms for the
speed-scaling case and analyze their computational performance. We also show
how to compute optimal schedules for the preemptive version of the problem in
polynomial time.
We investigate the impact of discussions on the quality of Crowdsourced content
using a data science approach that involves conducting an exploratory study to
uncover the associations among different discussion characteristics and article
quality and building a prediction model. By identifying appropriate instruments to
overcome selection, we build a model to quantify the impact of these
characteristics. Our results show that most of these characteristics have a positive
impact on quality.
285
TB06
2 - Toward Effective Information Diffusion on Social Media Platforms:
An Analysis of Dyadic Relationship
Jing Peng, The Wharton School, University of Pennsylvania, 3730
Walnut Street Suite 500, Philadelphia, PA, 19104, United States
of America, jingpeng@wharton.upenn.edu, Ashish Agarwal,
Kartik Hosanagar, Raghuram Iyengar
3 - Trading a Portfolio of Pairs in the Presence of Transaction Costs
James Primbs, Associate Professor, California State University
Fullerton, 800 N. State College Blvd., Fullerton, CA, United States
of America, jprimbs@fullerton.edu, Yuji Yamada
In this work we consider the problem of trading a portfolio of pairs when
transaction costs are present. We develop a receding horizon approach based on a
power utility function and proportional transaction costs. The resulting
methodology is very computational tractable, even for a portfolio of many
potentially correlated pairs. Backtested results on historical data are provided.
We investigate the impact of dyadic network characteristics on information
diffusion in social media platforms with directed networks. We propose a novel
hazard model to deal with the problem that a user may receive the information
from multiple others. The model is estimated using diffusion of ads on Digg.com.
We find that a non-reciprocal follower is more likely to adopt than a reciprocal
follower and the effects of network embeddedness are more complicated than
that in undirected networks.
4 - Backtesting Simultaneous Long-short and Proportional-integral
Investment Schemes
Sean Warnick, Associate Professor, Brigham Young University,
TMCB 2222, Provo, UT, 84602, United States of America,
sean.warnick@gmail.com, Scott Condie, Nathan Woodbury
3 - Monetizing Sharing Traffic through Incentive Design:
A Randomized Field Experiment
Tianshu Sun, University of Maryland Smith School of Business,
3330 Van Munching Hall, College Park, MD, 20740-2840,
United States of America, tianshusun@rhsmith.umd.edu,
Siva Viswanathan, Elena Zheleva
Simultaneous Long-Short is an investment strategy analyzed by Barmish and
Primbs that uses feedback control techniques to make investment decisions. An
extension of the technique uses proportional-integral control to make such
decisions. Importantly, these methods use a feedback architecture—and no
explicit market model—to manage investments. This study explores the
performance of these methods compared to other methods that use some estimate
of a market model through various backtests.
Customers share product information with each other everyday. While the share
of a product indicates clear purchase intent of either sender or recipient, most of
such sharing traffic does not lead to successful purchases. In collaboration with a
daily deal platform, we conduct a large field experiment to study whether and
how firms can monetize sharing traffic, by targeting senders with incentive.
Specifically, we examine the impact of incentive design on sender’s purchase as
well as referrals
■ TB07
4 - Founder and Funder, Just One Click Apart: How Social Media
Facilitates Investor Entrepreneur Match
Fujie Jin, the Wharton School, University of Pennsylvania, 500
Jon M Huntsman Hall, 3730 Walnut Street, Philadelphia, PA,
19104, United States of America, jinfujie@wharton.upenn.edu
Quantitative Risk Measurement and Modeling
Invited Session
Chair: Nan Chen, Professor, Chinese University of Hong Kong, 709A
William Mong Engineering Building, Hong Kong, Hong Kong - PRC,
nchen@se.cuhk.edu.hk
1 - On the Measurement of Economic Tail Risk
Xianhua Peng, Assistant Professor, Hong Kong University of
Science and Technology, Department of Mathematics,
Hong Kong, Hong Kong - PRC, maxhpeng@ust.hk
This study examines how entrepreneurs’ social media presence facilitates the
funding process across geographic regions. Comparison will be drawn between
traditional angel investors or VCs and the new crowdfunding platform to show
how entrepreneurs could optimally manage their social media profile to appeal to
different investor groups.
We show that the only risk measures that satisfy a set of economic axioms for the
Choquet expected utility and the statistical property of elicitability (i.e. there
exists an objective function such that minimizing the expected objective function
yields the risk measure) are the mean functional and the median shortfall, which
is the median of tail loss distribution. We argue that median shortfall is a better
alternative than expected shortfall for setting capital requirements in Basel
Accords.
■ TB06
Engineering Approaches in Finance
Sponsored Session
2 - Leverage, Market Liquidity, and Financial Fragility
Nan Chen, Prof, Chinese University of Hong Kong, 709A William
Mong Engineering Building, Hong Kong, Hong Kong - PRC,
nchen@se.cuhk.edu.hk, Jing Chen
Chair: James Primbs, Associate Professor, California State University
Fullerton, 800 N. State College Blvd., Fullerton, CA, United States of
America, jprimbs@fullerton.edu
1 - On Feedback Control-based Stock Trading: Some Back Tests
with High-frequency Data
B. Ross Barmish, Professor, University of Wisconsin, ECE
Department, Madison, WI, 53706, United States of America,
barmish@engr.wisc.edu
We provide a simple model to show how systemic fragility is built up as highly
leveraged investors crowd to similar trading strategies. As their wealth grows over
time, the destabilizing impact of their trading becomes more imminent, causing
amplified volatility, jump risk, and correlation co-movements in the security
prices.
3 - A Simulation Measure Approach to Monte Carlo Methods for
Default Timing Problems
Alex Shkolnik, University of California, Berkeley, CA, United
States of America, ads2@berkeley.edu, Kay Giesecke
The takeoff point for this paper is a new paradigm for stock trading involving
adaptive feedback control loops. I will first overview the key elements of our
theory with emphasis on “model-free” trading and money management.
Subsequently, I will describe recent back tests of our trading algorithms using
high-frequency data. Given that our underlying theory requires continuity of the
stock price, it is natural to study whether performance improves as a function of
the trading frequency.
Reduced-form models of name-by-name default timing are widely used to
measure portfolio credit risk and to analyze securities exposed to a portfolio of
names. Monte Carlo (MC) simulation is a common computational tool in such
settings. We introduce a new change of measure perspective for MC simulation
for default timing problems. The perspective provides the means of analyzing
current methods and suggests a new MC algorithm which outperforms a widely
used and standard technique.
2 - Construction of Nonlinear Simultaneous Equations Models for
Electricity Supply and Demand Functions
Yuji Yamada, Professor, University of Tsukuba,
3-29-1 Otsuka, Bunkyo-ku, Tokyo, 112-0012, Japan,
yuji@gssm.otsuka.tsukuba.ac.jp
In this work, we develop a new methodology for estimating supply and demand
functions in the Japan Electric Power Exchange (JEPX) spot market. To this end,
we generalize the standard simultaneous equations approach using linear
regressions for nonlinear case and show that the nonlinear structural equations
may be constructed based on the reduced equations of a nonparametric
regressions model. Then, we demonstrate the proposed approach using empirical
data.
286
■ TB08
TB10
2 - Network Visualization Analysis of Main Paths and Directions of
Firm Innovation
Jianxi Luo, Assistant Professor, Singapore University of
Technology and Design, 8 Somapah Rd, Singapore, 487372,
Singapore, luo@sutd.edu.sg, Bowen Yan
Sharing Economy and Peer-to-Peer Marketplaces
Invited Session
We present a method to represent the technology space as a network of patent
technology classes, and then overlay the network map to visualise firms’
technology capability positions and main paths of diversification over time. Based
on a few case studies, we show this method can reveal the differences in
innovation behaviours and strategies of different firms and aid in the assessment
of the firm’s past and existing capability positions and the exploration of future
innovation directions.
Chair: Jose Guajardo, University of California Berkeley,
545 Student Services Bldg #1900, Berkeley, CA, 94720-1900,
United States of America, jguajardo@berkeley.edu
1 - The Efficacy of Incentives in Scaling up Marketplaces
Ashish Kabra, INSEAD, Boulevard de Constance, Fontainebleau,
France, ashish.kabra@insead.edu, Elena Belavina, Karan Girotra
3 - Quantifying the Ecosystem of Digital Platform Companies
Rahul Basole, Associate Professor, Georgia Institute of
Technology, 85 Fifth Street NW, Atlanta, GA, 30332,
United States of America, basole@gatech.edu, Peter Evans
Achieving scale is key to the efficacy, survival and eventual domination of
marketplaces. Marketplace operators often run aggressive promotions and
incentive schemes to attract new users or increase the usage of existing users.
Using detailed transaction and location data from a leading transportation
marketplace, we estimate and compare the the short-term and long-term effects
of incentives given to the “buyer” side and “seller” side of the marketplace.
The rise of digital platforms is transforming industries and economies. Using an
integrated dataset (Crunchbase and Capital IQ), we quantify, compare, and
visualize the structure of 1,000+ platform companies. We discuss theoretical and
managerial implications.
2 - Business Models in the Sharing Economy: Manufacturing
Durables in Presence of Peer-to-peer Markets
Zhe Zhang, PhD Student, Carnegie Mellon University, 5000
Forbes Avenue, Pittsburgh, PA, 15213, United States of America,
zhezhang@cmu.edu, Jose Guajardo, Vibhanshu Abhishek
4 - Visualizing the Start-Up Genome
Raul Chao, ChaoR@darden.virginia.edu, Rahul Basole
Our study uses novel visual analytic techniques to analyze start-up activities.
Specifically, drawing on an analogy from genetics, we aim to visualize what we
refer to as the “Start-Up Genome” – a unique sequence of activities that defines
and differentiates one start-up from another.
We investigate the interaction between a manufacturer of durable goods and a
peer-to-peer marketplace where consumers trade the temporary use of the
durable good as a service. We analyze market outcomes under alternative
business models and market structures.
5 - Data-driven Visualizations of Market Differentiation in
Emerging Sectors
Martha Russell, Executive Director, mediaX at Stanford
University, 210 Panama Street, Cordura Hall,
Stanford, CA, 94305-4115, United States of America,
martha.russell@stanford.edu, Jukka Huhtamäki,
Neil Rubens, Kaisa Still
3 - Outsourcing Tasks Online: Matching Supply and Demand on
Peer-to-peer Internet Platforms
Chiara Farronato, Harvard Business School, Soldiers Field,
Boston, United States of America, chiarafarronato@gmail.com,
Zoe Cullen
Using keywords from a dataset built from online promotional information, we
visualize the character and strength of startups’ market objectives in emerging
sectors.
We study a central problem for peer-to-peer markets: how to create matches
when demand and supply are highly variable. We develop a model of a matching
market for services, and estimate it using data from TaskRabbit. We find that
supply is highly elastic and estimate average gains from each trade to be $37.
Because of the matching frictions, the ex-ante gains are more modest, but are
maximized by the elastic supply. Finally, we explore heterogeneity of platform
success across cities.
■ TB10
4 - First Ranked First to Serve: Strategic Agents in a Service Contest
Konstantinos Stouras, PhD Candidate, INSEAD,
Bd. de Constance, Fontainebleau, 77305, France,
Konstantinos.STOURAS@insead.edu, Karan Girotra,
Serguei Netessine
Frontiers in IS Research
Sponsor: E-Business
Sponsored Session
Chair: Min-Seok Pang, Assistant Professor, Temple University, 1810 N
13th St, Speakman 201e, Philadelphia, PA, 19122, United States of
America, minspang@temple.edu
1 - Fundraising Patterns and Entrepreneurial Performance in
Crowdfunding Platforms
Eun Ju Jung, George Mason University, Enterprise Hall, 4400
University Drive, School of Business, Fairfax, VA, 22030, United
States of America, jej978@gmail.com, Vallabh Sambamurthy,
Anjana Susarla
We develop a model of a virtual call center that pays its agents on-demand, by
committing to a (relative) performance agent ranking prioritization scheme. We
show that the optimal design of such a “service contest” is often coarse.
Discarding available information about agents’ relative performance, or deploying
coarser priority classes can paradoxically create higher incentives for agents to
voluntarily participate and provide better service.
■ TB09
Crowdfunding provides entrepreneurs with new opportunities for funding and
ultimately fosters entrepreneurship and new firm creation. However, there is a
dearth of research on entrepreneurial performance after fundraising success. In
this paper, we examine how the dynamics in fundraising processes are related to
entrepreneurial performance.This study will contribute to crowdfunding and
entrepreneurship literature and offer practical implications.
Ecosystem Analytics & Visualization
Sponsored Session
Chair: Rahul Basole, Associate Professor, Georgia Institute of
Technology, 85 Fifth Street NW, Atlanta, GA, 30332,
United States of America, basole@gatech.edu
1 - Integrated Analytics Framework for Business
Ecosystem Dynamics
Hyunwoo Park, Georgia Institute of Technology, 85 5th St. NW
Rm 339, Atlanta, GA, 30309, United States of America,
hwpark@gatech.edu, Rahul Basole
2 - It Security Effectiveness: Influence of Breach Type and
Public Attention
John D’arcy, University of Delaware, 207A Purnell Hall,
Newark, DE, 19716, United States of America, jdarcy@udel.edu,
Asli Basoglu
This study explores factors that bias auditor judgments of companies’ information
security effectiveness. We developed a dataset consisting of security breaches
against publicly traded companies, public attention attributed to these breaches–in
the form of abnormal Google search volume, and auditor evaluations of these
companies’ IT internal controls. Our results suggest that breach source and
abnormal public attention both contribute to biased evaluations of information
security effectiveness.
We propose a computational framework and interactive prototype for specifying
and analyzing business ecosystem dynamics. Our research fuses simulation with
data/process mining and information visualization techniques, enabling decision
makers to specify micro-behavior of firms, generate and test hypotheses, gain
insights, and communicate results effectively. We illustrate our approach using
real-world examples based on a unique curated dataset from multiple sources.
287
TB11
■ TB12
3 - Studying Influence of Comments in Online News Papers
Iljoo Kim, Assistant Professor, Saint Joseph’s University, 347
Mandeville Hall, 5600 City Avenue, Philadelphia, PA, 19131,
United States of America, ikim@sju.edu, Gautam Pant
Nonlinear Programming in Stochastic and
Multilevel Problems
In this work, we study online comments and their influence in online news
articles. Using text-mining techniques, we attempt to explain and/or predict
influence of online newspaper comments on the context of the original article or
even on creating a new agenda through the discussions among commenters. This
is done based on the textual signals embedded within comments as well as news
articles.
Sponsor: Optimization/Mixed Integer Nonlinear Optimization and
Global Optimization
Sponsored Session
Chair: Alexander Vinel, Auburn University, 3301 Shelby Center,
Auburn, AL, 36849-5346, United States of America,
alexander.vinel@auburn.edu
1 - Branch-and-cut Algorithm for Integer Bilevel Linear
Optimization Problems
Sahar Tahernejad, Graduate Student, Lehigh University, 12 Duh
Drive- No. 132, Bethlehem, PA, 18015, United States of America,
sat214@lehigh.edu, Ted Ralphs
4 - Politics and Information Technology Investments in
The U.S. Federal Government in 2003-2015
Min-Seok Pang, Assistant Professor, Temple University,
1810 N 13th St, Speakman 201e, Philadelphia, PA, 19122,
United States of America, minspang@temple.edu
What makes some US federal agencies digitally advanced and others lagging? This
study investigates how politics affects IT investment in federal agencies. With a
panel dataset from 133 federal agencies, our empirical analyses produce several
intriguing findings. A federal agency makes more capacity-building IT
investments (i) when its head is appointed with legislative approval, (ii) when the
federal government is less divided, and (iii) when it is neither too conservative
nor too liberal.
We extend the branch-and-cut framework of Denegre and Ralphs for solving
integer bilevel linear optimization problems (IBLPs). IBLPs differ from standard
integer optimization problems in that there are solutions which are integer but
not feasible and they should be removed from the feasible solution set. Our
proposed algorithm applies a variety of cut generation techniques for removing
such solutions. We report on numerical experiments on some benchmark IBLPs.
■ TB11
2 - On Pessimistic Versus Optimistic Bilevel Linear Programs
M. Hosein Zare, University of Pittsburgh, 1048 Benedum Hall,
Pittsburgh, PA, 15261, United States of America, moz3@pitt.edu,
Osman Ozaltin, Oleg Prokopyev
Machine Learning under a Modern Optimization Lens
Sponsored Session
We study the relationships between Pessimistic and Optimistic Bilevel Linear
Programs. In particular, we focus on the case when the upper-level decisionmaker (i.e., the leader) needs to consider the uncertain behavior of the
lower-level decision maker (i.e., the follower). We derive some computational
complexity properties, and illustrate our results using a defender-attacker
application.
Chair: Dimitris Bertsimas, Professor, MIT, 77 Massachusetts Ave.,
Cambridge, MA, 02139, United States of America, dbertsim@mit.edu
1 - Sparse Principal Component Analysis via a Modern
Optimization Lens
Lauren Berk, Massachusetts Institute of Technology, 77
Massachusetts Avenue, Bldg. E40-149, Cambridge, MA, 02139,
United States of America, lberk@mit.edu, Dimitris Bertsimas
3 - Identifying Risk-averse Low-diameter Clusters in Graphs with
Random Vertex Weights
Maciej Rysz, NRC-AFRL, 1350 N. Poquito Road, Shalimar, FL,
United States of America, mwrysz@yahoo.com, Pavlo Krokhmal
We consider the problem of finding a k-club of minimum risk contained in a
graph whose vertices have stochastic weights. A stochastic programming
framework that is based on the formalism of coherent risk measures is used to
find the corresponding subgraphs. A combinatorial branch-and-bound solution
algorithm is proposed.
We develop tractable algorithms that provide provably optimal solutions to the
exact Sparse Principal Component problems of up to 1000 dimensions, using
techniques from Mixed Integer Optimization and first order methods. Unlike
earlier SPCA methods, our approach retains complete control over the degree of
sparsity of the components, and provides solutions with higher explained
variance.
2 - Robust Support Vector Machines
Colin Pawlowski, MIT, 77 Massachusetts Ave., Cambridge, MA,
02139, United States of America, cpawlows@mit.edu,
Dimitris Bertsimas
4 - Solution Procedures for a Class of Mixed-integer Nonlinear
Programming Problems
Alexander Vinel, Auburn University, 3301 Shelby Center,
Auburn, AL, 36849-5346, United States of America,
alexander.vinel@auburn.edu, Pavlo Krokhmal
We consider a maximal-margin classifier which is the non-regularized
formulation of SVM. Using Robust Optimization, we develop new,
computationally tractable methods that are immunized against uncertainty in the
features and labels of the training data. Experiments on real-world datasets from
the UCI Machine Learning Repository show out-of-sample accuracy
improvements for robust methods in a significant number of problems analyzed.
We study solution approaches for a class of mixed-integer non-linear
programming problems with our interest stemming from recent developments in
risk-averse stochastic programming. We explore possible applications of some of
the solution techniques that have been successfully used in mixed-integer
second-order conic programming and show how special structure of problems
under consideration can be utilized.
3 - Optimal Trees
Jack Dunn, Operations Research Center, MIT, 77 Mass Ave,
Bldg E40-130, Cambridge, MA, 02139, United States of America,
jackdunn@mit.edu, Dimitris Bertsimas
■ TB13
Decision trees are widely used to solve the classical statistical problem of
classification. We introduce a new method for constructing optimal decision trees
using Mixed-Integer Optimization, and show using real data sets that these trees
can offer significant increases in accuracy over current state-of-the-art decision
tree methods. We also demonstrate the benefits of using Robust Optimization
when constructing these trees.
Stochastic Approximation
Sponsored Session
Chair: Raghu Pasupathy, Associate Professor, Department of Statistics,
Purdue University, 250 N University Street, West Lafayette, IN, 47907,
United States of America, pasupath@purdue.edu
1 - Budget-constrained Stochastic Approximation
Uday Shanbhag, The Pennsylvania State University, 310
Leonhard Building, University Park, PA, 16801, United States of
America, udaybag@engr.psu.edu, Jose Blanchet
4 - Logistic Regression using Robust Optimization
Daisy Zhuo, Massachusetts Institute of Technology, 77
Massachusetts Avenue, Cambridge, MA, 02139,
United States of America, zhuo@mit.edu, Dimitris Bertsimas
Logistic regression is one of the most commonly used classification methods, yet
the solution can be sensitive to inaccuracy and noise in data. Here we propose an
approach using Robust Optimization to find stable solutions under uncertainties
in data features and labels. Using more than 80 real-world problems, we
demonstrate that the robust logistic regressions lower misclassification error
significantly in the majority of the data sets.
We consider a convex constrained stochastic convex optimization problem in
which the simulation budget is fixed and computation is expensive. We consider
stochastic approximation schemes in which the sample-size is either constant or
updated at every step while meeting this budget and provide suitable finite-time
error bounds.
288
TB16
■ TB15
2 - Optimal Averaging Schemes for Stochastic
Approximation Methods
Farzad Yousefian, Postdoctoral Research Associate, Penn State,
333 Logan Ave., Apt. 307, State College, PA, 16801, United States
of America, szy5@psu.edu, Angelia Nedich, Uday Shanbhag
Recent Advances in Nonlinear Programming
Sponsor: Optimization/Nonlinear Programming
Sponsored Session
We develop optimal weighted averaging stochastic approximation schemes for
solving stochastic variational inequality problems. We show that the gap function
associated with the averaged sequence diminishes to zero at the optimal rate. We
also develop a window-based variant of this scheme that displays the optimal rate
and the superiority in the constant factor of the bound comparing to the classic
averaging schemes. Preliminary numerical results on a stochastic Nash-Cournot
game are presented.
Chair: Hande Benson, Associate Professor, Drexel University, LeBow
College of Business, Philadelphia, PA, 19104, United States of America,
hvb22@drexel.edu
1 - Solving the Problem of Portfolio VAR Minimization as a Nonlinear
Program
Arun Sen, Director, Navigant Consulting, 685 3rd Avenue,
14th Floor, New York, NY, 10017, United States of America,
arunsen@alumni.princeton.edu
3 - Adaptive Sampling Line Search for Local Simulation Optimization
Raghu Pasupathy, Associate Professor, Department of Statistics,
Fatemeh Hashemi
Minimizing Value at Risk (VAR) is challenging as the optimization problem is
non-convex. In previous work the problem was formulated as an MPEC
(mathematical program with equilibrium constraints), that was solved using
branch-and-bound techniques. We show that the same MPEC can be solved
effectively as a nonlinear program, specifically by use of interior-point methods,
and that this a flexible approach that is easily able to incorporate additional
constraints on the optimal portfolio.
We present an algorithm for continuous simulation optimization that combines
adaptive sampling ideas with a classical line search method from deterministic
nonlinear programming. We will discuss theoretical properties and a brief
example.
4 - Noisy Collective Nonconvex Optimization
Mengdi Wang, Assistant Professor, Princeton University, 302
Trinity Ct #2, Princeton, NJ, 08540, United States of America,
mengdiw@princeton.edu
2 - Fast Algorithms for LAD Lasso Problems
Robert Vanderbei, Princeton University, ORFE,
Sherrerd Hall, Princeton, NJ, 08544,
United States of America, rvdb@princeton.edu
Paper not available at this time.
We will present a new algorithm for the LAD Lasso problem. We will compare
this new algorithm against existing state-of-the-art algorithms.
■ TB14
3 - Cubic Regularization for First-order Methods
David Shanno, RUTCOR - Rutgers University (Emeritus),
Rutgers University, New Brunswick, NJ, United States of
America, shannod@comcast.net, Hande Benson
Joint Session OPT/ICS: Stochastic Programming:
Progressive Hedging and Related Methods
Regularization techniques have been used to help existing algorithms solve
“difficult” nonlinear optimization problems. Over the last decade, regularization
has been proposed to remedy issues with equality constraints and equilibrium
constraints, bound Lagrange multipliers, and identify infeasible problems. In this
talk, we will focus on the application of cubic regularization in the context of the
symmetric rank one and the conjugate gradient methods for nonlinear
programming.
Sponsored Session
Chair: Jonathan Eckstein, Professor, Rutgers University, 100 Rockafeller
Road, Piscataway, NJ, 08854, United States of America,
jeckstei@rci.rutgers.edu
1 - Scalable Lower and Upper Bounding Techniques for Stochastic
Unit Commitment with Progressive Hedging
Jean-paul Watson, Sandia National Laboratories,
P.O. Box 5800, MS 1326, Albuquerque, United States of America,
jwatson@sandia.gov, David Woodruff, Sarah Ryan
4 - Value Driven Design Delegation - An Optimization Model
Vinod Cheriyan, Enova International, 1255 S Michigan Ave.
Apt 3711, Chicago, IL, 60605, United States of America,
vinod.cheriyan@gmail.com, Chris Paredis, Anton Kleywegt
Rather than satisfaction of stakeholder needs, the Value Driven Design approach
focuses on maximization of economic value. For large, complex systems, the
systems designer maximizes the value by delegating detailed design to many
subsystem teams. We study the convergence properties of such a value-driven,
delegation-based system design process, where knowledge is distributed. We
model the design as an optimization problem. We propose an algorithm and show
that it converges to a critical point.
We describe configurations of a scenario-based decomposition strategy for solving
the stochastic unit commitment problem, based on the progressive hedging
algorithm. We consider both upper and lower bounding aspects of progressive
hedging in the mixed-integer case, and demonstrate parameterizations that yield
extremely tight optimality gaps for 100-generator cases and moderately tight
optimality gaps for 350-generator cases.
2 - Progressive Hedging and Dual Decomposition
David Woodruff, UC Davis, One Shields Avenue, Davis, CA,
95616, United States of America, dlwoodruff@ucdavis.edu
■ TB16
The PH algorithm proposed by Rockafellar and Wets and the DDSIP algorithm
proposed by Caroe and Schultz can both be thought of as primal-dual algorithms
and both can be used to address stochastic mixed-integer programs. In this talk I
describe work with numerous co-authors to use the two algorithms together. In
addition we describe an algebraic modeling language (Pyomo) interface to DDSIP
that is useful with, or without, PH.
3 - Asynchronous Projective Progressive-hedging-like Stochastic
Programming Decomposition Methods
Jonathan Eckstein, Professor, Rutgers University, 100 Rockafeller
Road, Piscataway, NJ, 08854, United States of America,
jeckstei@rci.rutgers.edu
Chair: Sertalp Cay, Lehigh University, 200 W Packer Ave, Bethlehem,
PA, 18015, United States of America, sec312@lehigh.edu
1 - Portfolio Optimization Problems with Cone Constraints and
Discrete Decisions
Umit Saglam, Assistant Professor, East Tennessee State University,
Department of Management and Marketing, College of Business
and Technology, Johnson City, TN, 37614, United States of
America, saglam@etsu.edu, Hande Benson
Various Aspects of Mixed Integer Conic Optimization
Sponsored Session
We present a class of stochastic programming algorithms based on new
Combettes-Eckstein monotone operator splitting methods. Unusually, these
splitting methods need to re-solve only a subset of the subproblems at each
iteration, using boundedly outdated information. Applying these techniques to
stochastic programming yields methods that resemble progressive hedging, but
can operate in a fully asynchronous manner. Convergence is guaranteed under
the same conditions as for progressive hedging.
In this study we consider both single-period and multiperiod portfolio
optimization problems based on the Markowitz (1952) mean/variance
framework. Our model is aggregated from current literature.We solve these
models with a MATLAB based Mixed Integer Linear and Nonlinear Optimizer
(MILANO). We have devised and implemented the first solution method for such
problems and demonstrate its efficiency on large-scale portfolio optimization
models.We also provide substantial improvement in runtimes.
289
TB17
2 - Computational Study of a Second Order Cone Relaxation for
Binary Quadratic Polynomial Problems
Julio Goez, Postdoctoral Fellow, Ecole Polytechnique Montreal
and GERAD, 2900 Boulevard Edouard-Montpetit, Montréal, QC,
H3T 1J4, Canada, jgoez1@gmail.com, Miguel Anjos
4 - Network Design Problem for Battery Electric Bus
Yousef Maknoon, EPFL, Route Cantonale, Lausanne, Switzerland,
yousef.maknoon@epfl.ch, Shadi Sharif Azadeh, Michel Bierlaire
In electric bus planning, for battery installation, we need to investigate two
points: (1) the type and location of charger stations (2) the capacity of battery of
each bus. In this presentation, first we describe the problem and the design
elements. Then, we present its mathematical form followed by the resolution
approach. Finally, we demonstrate the computational results on our case study
and discuss about the robustness of the plan.
This work presents a computational study of the second order cone relaxation for
binary quadratic problems proposed by Ghaddar, Vera and Anjos (2011) who
used a polynomial optimization approach. We explore how this relaxation can be
strengthened using additional constraints, and also, we explore the relation of
disjunctive conic cuts with this relaxation.
3 - Computational Approaches to Mixed Integer Second Order
Cone Optimization
Aykut Bulut, PhD Candidate, Lehigh University, 200 W. Packer
Ave., Bethlehem, PA, 18015, United States of America,
aykut@lehigh.edu, Ted Ralphs
■ TB18
Data Mining for Different Type of Big Data
Invited Session
We introduce an open-source Mixed Integer Second Order Cone Optimization
(MISOCO) solver. We present computational experiments on various approaches
to solve MISOCO problem. We test using outer approximation method to solve
continuous relaxations. We also discuss using various valid inequalities to
improve the continuous relaxations. We discuss computational performance of
these approaches on conic benchmark library (CBLIB 2014) problems.
Chair: Young-seon Jeong, Chonnam National University,
Department of Industrial Engineering, Gwangju, Korea, Republic of,
youngseonjeong@gmail.com
1 - Classification of Uncertain Data using Group to Object Distances
Behnam Tavakkol, PhD Candidate, Rutgers University,
5200 BPO Way, Piscataway, NJ, 08854, United States of America,
btavakkol66@gmail.com, Myong K (MK) Jeong, Susan Albin
4 - Solving Robust Portfolio Optimization Problems in Practice
Sarah Drewes, Senior Consultant, Dr., MathWorks, Adalperostr.
45, Ismaning, Germany, Sarah.Drewes@mathworks.de
Robust versions of the Markowitz mean-variance model can reduce the
estimation risk induced by its sensitivity to changes in expected returns or the
covariance matrix. Probabilistic versions of the classical model can be formulated
as nonlinear and often second order cone programs. We study how to solve these
problems also by general nonlinear solvers (MATLAB Optimization Toolbox) and
in case of discrete variables. We evaluate both computational performance and
complexity of implementation.
Uncertain data problems have features represented by multiple observations or
their fitted PDFs. We propose two approaches for classifying uncertain data
objects. The first uses existing Probabilistic Distance Measures for object-to-object
distances and classifies with KNN. The second features a new probabilistic
distance measure for object to class distances.
2 - The Classification Methodology of Chip Quality using Canonical
Correlation Analysis
Ki-hyun Kim, Samsung Electronics Co., Banwol-dong,
Hwaseong-si, Gyeonggi-do, Korea, Republic of,
bluenamja@daum.net
■ TB17
In this study, we proposed classification methodology using a canonical
correlation analysis as feature selection method at multi-dimensional chip level
data generated in the semiconductor manufacturing industry. As the result of this
research, we were able to extract important varialbes in the varous PCM variables
from the correlation of the multiple FBC variables and PCM variables. The
proposed method was improved the accuracy of quality classification for a chip
tested in the probe test.
Network Modeling and Design
Sponsored Session
Chair: Mario Ventresca, Assistant Professor, School of Industrial
Engineering, Purdue University, 315 N Grant St, Lafayette, IN, 47906,
United States of America, mventresca@purdue.edu
1 - Action-based Network Models
Viplove Arora, Graduate Student, School of Industrial
Engineering, Purdue University, 45 N Salisbury St, Apt. 9,
West Lafayette, IN, 47906, United States of America,
arora34@purdue.edu, Mario Ventresca
3 - Multivariate Monitoring for Metal Fabrication Process in Mobile
Devices Manufacturing
Seonghyeon Kang, M.s. Candidate, Korea University, Innovation
Hall 817, Korea University, 145 Anam-ro, Seongbuk-gu, Seoul,
136-713, Korea, Republic of, shyeon.kang@gmail.com,
Seoung Bum Kim
Complex networks are very useful representations of real world complex systems.
A number of network generation procedures have been proposed that are capable
of producing networks with a restricted subset of structural properties. However,
a unifying model remains elusive. We present initial results on an action-based
perspective that has potential to yield more general network structures than
existing techniques. A machine learning approach to learn a probabilistic model
will be presented.
In mobile industry, using metal case of devices is rapidly increased for thin and
attractive design. However, fabricating metal is the difficult process because
accurate control of equipment are required. In this study, we propose an efficient
multivariate monitoring procedure to observe more than 40 parameters of metal
fabrication equipment. The effectiveness of the proposed procedure is
demonstrated by real data from the mobile plant in one of the leading mobile
companies in South Korea.
2 - Automatically Inferring Complex Network Models
Mario Ventresca, Assistant Professor, School of Industrial
Engineering, Purdue University, 315 N Grant St., Lafayette, IN,
47906, United States of America, mventresca@purdue.edu
4 - Multivariate Monitoring of Automated Material Handling Systems
in Semiconductor Manufacturing
Sangmin Lee, Korea University, Innovation Hall 817, Korea
University, 145 Anam-ro, Seongbuk-gu, Seoul, 136-713, Korea,
Republic of, smlee5679@gmail.com, Seoung Bum Kim
Complex networks are becoming increasingly important across many disciplines.
However, the problem of network modeling is extremely intricate and time
consuming. Hence, frameworks have been proposed to estimate model
parameters, but are focused on capturing a small subset predefined network
characteristics such as degree distribution. I will present recent work on a highly
robust automated inference approach that is able to discover arbitrary network
models with minimal human insight.
Monitoring all possible contingencies in automated material handling system
(AMHS) of semiconductor manufacturing is a difficult task because tremendous
hardware and software systems are involved. This study presents an efficient
multivariate monitoring procedure to monitor more than 100 KPIs in AMHS. The
effectiveness and applicability of the proposed procedure is demonstrated by real
data from semiconductor fabrication plant in one of the leading semiconductor
companies in South Korea.
3 - Cut-set Separation Schemes for the Robust Single-commodity
Network Design Problem
Daniel Schmidt, Carnegie Mellon University, 720 S Negley Ave,
schmidtd@cmu.edu, Chrysanthos Gounaris
5 - Quantifying the Level of Risk of Functional Chips in
Semiconductor Wafers
Young-seon Jeong, Chonnam National University, Department of
Industrial Engineering, Gwangju, Korea, Republic of,
youngseonjeong@gmail.com, Byunghoon Kim,
Seung-hoon Tong, Inkap Chang, Myong K (MK) Jeong
We address the exact solution of the Robust Single-Commodity Network Design
Problem in which customer demands are uncertain and realize from within
anappropriately defined uncertainty polytope. We explore techniques to
approximate the arc-flow based formulation with fewer variables. We also
evaluate the use of meta-heuristics for the NP-hard problem of separating cut-set
inequalities in the context of a branch-and-cut solution approach.
This talk presents the procedure to quantify the level of risk of functional chips in
dynamic random access memory (DRAM) wafers. To screen risky functional
chips, the risk level of each chip is estimated by the posterior probability for
functional chips. The functional chips closer to the class of defective chips may
have a higher probability of being failed in the near future. The experimental
results by using real-life wafers show the effectiveness of the proposed method.
290
■ TB21
TB23
2 - Statistical Guarantees for Individualized Rank Aggregation
Sahand Negahban, Yale University, 24 Hillhouse Ave,
New Haven, CT, 06510, United States of America,
sahand.negahban@yale.edu
Bundled Payment Systems
We study a version of rank aggregation known as collaborative ranking. In this
problem we assume that individual users provide us with pairwise preferences
and from those preferences we wish to obtain rankings on items that the users
have not had an opportunity to explore. We provide a theoretical justification for
a nuclear norm regularized optimization procedure.
Sponsored Session
Chair: Danny Hughes, Harvey L. Neiman Health Policy Institute,
1891 Preston White Drive, Reston, VA, 20191,
United States of America, dhughes@neimanhpi.org
1 - Optimizing Provider Decisions under Bundled Payments
Brenda Courtad, University of Cincinnati, 2925 Campus Green
Dr, Cincinnati, OH, 45221, United States of America,
courtabl@mail.uc.edu
3 - Inference in High-dimensional Varying Coefficient Models
Mladen Kolar, Assistant Professor, Chicago Booth, 5807 South
Woodlawn Avenue, Chicago, IL, 60637, United States of America,
mkolar@chicagobooth.edu, Damian Kozbur
We focus on the high-dimensional linear varying-coefficient model and develop a
novel procedure for estimating the coefficient functions. Our procedure works in
a high-dimensional regime, under arbitrary heteroscedasticity in residuals, and is
robust to model misspecification. We derive an asymptotic distribution for the
normalized maximum deviation of the estimated coefficient function and
demonstrate how these results can be used to make inference in highdimensional dynamic graphical models.
When moving from fee-for-service to bundled payments the providers focus shifts
from revenue generating to cost reducing. We develop a partially observable
Markov decision process to aid providers in deciding which interventions to
implement to reduce costs.
2 - Increasing Healthcare Value in Accountable Care Organizations
through Incentive Redesign
Christian Wernz, Virginia Tech, Industrial and Systems
Engineering, Blacksburg, VA, 24060, United States of America,
cwernz@vt.edu, Hui Zhang, Barry Barrios, Danny Hughes
4 - Elementary Estimators for High-dimensional Statistical Models
Eunho Yang, IBM T.J. Watson, P.O. Box 218, Yorktown Heights,
United States of America, eunho@cs.utexas.edu
We propose a class of closed-form estimators for sparsity-structured highdimensional models. Our approach builds on observing the precise manner in
which the classical MLE breaks down under high-dimensional settings. We
provide a rigorous statistical analysis that shows that our simple estimators
recovers the same asymptotic convergence rates as those of computationally
expensive L1-regularized MLEs. We corroborate statistical performance, as well as
computational advantages via simulations.
ACOs are incentivized by the Centers for Medicare and Medicaid Services (CMS)
to lower costs while meeting quality standards. We determined how CMS’
incentive program can be improved, and how ACOs can optimally distribute
incentives among their members. Using multiscale decision theory, we performed
an in-depth analysis of computed tomography (CT) scanner investments and use
in ACO hospitals, calibrated the model with Medicare data, and show how CT
scan cost can be lowered and outcomes improved.
3 - Mitigating Small Provider Financial Risk under Prospective
Bundled Payment Systems
Danny Hughes, Harvey L. Neiman Health Policy Institute,
1891 Preston White Drive, Reston, VA, 20191, United States of
America, dhughes@neimanhpi.org, Jeremy Eckhause
■ TB23
New Advances in Production Planning and
Scheduling
Retrospective bundled payment models, which cover all medical services within
an episode of care, usually include stop loss provisions to manage financial risk.
As payments shift to prospective bundled payments, the mechanisms to manage
these stop loss provisions may no longer exist. We develop nonlinear
programming models to develop pricing strategies that address the inherent
higher risk to smaller providers under such a payment system.
Cluster: Stochastic Models: Theory and Applications
Invited Session
Chair: Jingshan Li, Professor, 1513 University Ave, Madison, WI,
53706, United States of America, jli252@wisc.edu
1 - Coordination in Multi-product Manufacturing Systems: Modeling
and Analysis
Cong Zhao, Research Assistant, University of Wisconsin-Madison,
1513 University Ave, Room 3235, Madison, Wi, 53706,
United States of America, czhao27@wisc.edu, Ningxuan Kang,
Li Zheng, Jingshan Li
4 - Bundled Payments: The Roles of Organization and Diagnosis
Turgay Ayer, Georgia Institute of Technology, 765 Ferst Dr.,
Atlanta, GA, 30332, United States of America,
ayer@isye.gatech.edu, Mehmet Ayvaci, Jan Vlachy
Medicare has started contracting with healthcare providers for bundled payments.
However, most providers do not have experience with the risks and opportunities
for such payment mechanism. We propose a game-theoretic model to capture the
power dynamics between physicians and the hospital under various patient
pathways. We use the model to generate hypotheses and test these hypotheses
using real data.
Multi-product systems are common in today’s manufacturing process. Effective
coordination between products in such systems is important in operation. We
study a two-product geometric manufacturing system and derive closed-form
expressions of performance measures. An optimal allocation policy of buffer
thresholds is developed, and the monotonicity of optimal buffer size with respect
to machine parameters is investigated. The managerial insights to achieve optimal
production control are discussed.
■ TB22
2 - Level Scheduling in Automotive Assembly Lines and its Effect on
the Consumption of Ressources
Heinrich Kuhn, Professor, Catholic University of EichstaettIngolstadt, Supply Chain Management & Operations,
Auf der Schanz 49, Ingolstadt, 85049, Germany,
heinrich.kuhn@ku-eichstaett.de, Dominik Woerner
Learning High-dimensional/ Sparse Models
Sponsored Session
Chair: Varun Gupta, University of Chicago Booth School of Business,
Chicago, IL, United States of AmericaVarun.Gupta@chicagobooth.edu
1 - Robust Methods for High-dimensional Regression
Po-ling Loh, Assistant Professor, University of Pennsylvania, 3730
Walnut St, 466 Jon M. Huntsman Hall, Philadelphia, PA, 19104,
United States of America, loh@wharton.upenn.edu
Level scheduling approaches in sequencing of assembly lines are used as
substitutional model for the underlying economic and sustainable objectives since
a leveled distribution of materials requirements does not necessarily contribute
directly to these objectives. We conduct a case study at a major German
automotive company selecting relevant part families whose consumption is
currently unequal distributed by an extensive simulation study.
We discuss new methods for robust regression in high-dimensional settings. Our
procedures draw upon classical approaches in robust statistics, designed for
scenarios where the number of parameters is fixed and the sample size grows to
infinity — however, these methods may be adapted to perform robust inference
in high dimensions, as well. We also prove that the robust estimators, which
involve minimizing nonconvex functions, may nonetheless be optimized to
desirable accuracy.
3 - A New Scenario Based Sales and Operations Planning Model
Nico Vandaele, Professor, KU Leuven, Naamsestraat 69 Box 3555,
Leuven, 3000, Belgium, nico.vandaele@kuleuven.be,
Catherine Decouttere, Gerd Hahn, Torben Sens
We apply a scenario-based approach to the sales and operations planning process
where both model-based and non-model based Key Performance Indicators are
taken into account. This allows us to balance customer service, derived from
aggregate order lead times, and relevant costs of operations when determining
volume/mix decisions for internal and external production. An industry-derived
case example with distinct outsourcing options is used to highlight the benefits of
the approach.
291
TB24
■ TB24
concepts of ‘digital borders’ and ‘border strength’ and use them in an
experimental investigation.
3 - Design Control in Open Innovation: An Examination of Open
Source Software Production
Shivendu Pratap Singh, University of Pittsburgh, Room 229,
Mervis Hall, Pittsburgh, 15260, United States of America,
shs161@pitt.edu
Urban Data Analytics and Mining
Sponsored Session
Chair: Xun Zhou, Assistant Professor, University of Iowa, S210
Papajohn Business Building, 21 E Market Street, Iowa City, IA, 52242,
United States of America, xun-zhou@uiowa.edu
1 - A Data Mining Approach to the Discovery of Emerging Hotspot
Patterns in Urban Data
Amin Vahedian Khezerlou, University of Iowa, S283
Pappajohn Business Building, The University of Iowa,
Iowa City, IA, 52242-1994, United States of America,
amin-vahediankhezerlou@uiowa.edu, Xun Zhou
Firms are opting for co-creating software, by attracting developers on platforms
like GitHub.This shared model of development requires flexible software design
controls to influence community engagement, which could result in proliferation
of design options. Flexible design control policy could have side effects such as
accumulation of technical debt and need to be judiciously managed. This paper
examines the antecedents and consequences of design control policies in social
software production.
4 - Time-dependent Pricing for Mobile Data: Analysis, Systems,
and Trial
Soumya Sen, ssen@umn.edu, Carlee Joe-wong, Mung Chiang,
Sangtae Ha
Emerging hotspots can be observed in urban data, e.g., cellular service or traffic
congestions due to non-periodic events (e.g., sport games, accidents). Efficiently
identifying these patterns help city planners and service provides response early
to the congestions. Previous hotspot detection techniques focused on patterns
with fixed footprints. We propose an efficient data mining approach to detect
emerging congestion patterns with dynamic footprints and validate our method
on real urban data.
Dynamic pricing of mobile data traffic can alleviate network congestion by
creating temporally-varying price discounts. But realizing it requires developing
analytical models for price point computation, systems design, and field
experiments to study user behavior. In this paper, we present the architecture,
implementation, and a user trial of a day-ahead time-dependent pricing.
2 - Exploiting Geographic Dependencies for Real Estate Ranking
Yanjie Fu, Rutgers University, 504 N 5th St, Harrison, NJ, 07029,
United States of America, yanjie.fu@rutgers.edu, Hui Xiong,
Hui Xiong
■ TB26
We propose a geographic method, named ClusRanking, for estate evaluation by
leveraging the mutual enforcement of ranking and clustering power model the
geographic dependencies of estates for enhancing estate ranking. Indeed, the
geographic dependencies of the investment value of an estate can be from the
characteristics of its own neighborhood (individual dependency), the values of its
nearby estates (peer dependency), and the prosperity of the affiliated latent
business area (zone dependency).
Retailer Pricing
Invited Session
Chair: Kathy Stecke, UT Dallas, SM30 JSOM, 800 W Campbell Rd,
Richardson, TX, 75080, United States of America, kstecke@utdallas.edu
3 - A General Geographical Probabilistic Factor Model for Point of
Interest Recommendation
Bin Liu, Rutgers University, 900 Davidson Road, 47 Nichols
Apartment, Piscataway, NJ, 08854, United States of America,
binben.liu@rutgers.edu
Co-Chair: Xuying Zhao, University of Notre Dame, Notre Dame, IN,
Xuying.Zhao.29@nd.edu
1 - Optimal Price Trajectories and Inventory Allocation for Inventory
Dependent Demand
Stephen Smith, Professor, Santa Clara University, 500 El Camino
Real, Lucas Hall 216H, Santa Clara, CA, 95053-0382, United
States of America, ssmith@scu.edu, Narendra Agrawal
The problem of point of interest recommendation is to provide personalized
places. The decision process for a user to choose a POI can be influenced by
numerous factors, such as personal preferences, geographical considerations, and
user mobility behaviors. We propose a general geographical probabilistic factor
model framework which takes various factors into consideration. Extensive
experimental results show promise of the proposed methods.
Retail demand is often inventory dependent because larger inventories create
more attractive displays and low inventories can create broken assortments. This
research jointly optimizes the price trajectory and the allocation of a given
amount of inventory across a set of non-identical stores with inventory
dependent demands.
■ TB25
2 - The Effect of Reward Purchase on Dynamic Pricing
Hakjin Chung, Stephen M. Ross School of Business,
University of Michigan, Ann Arbor, MI, United States of America,
hakjin@umich.edu, So Yeon Chun, Hyun-soo Ahn
Software-Driven Innovation and Business Strategies
In many loyalty programs, consumers are provided with an option to acquire
products by redeeming loyalty points instead of cash. We characterize when
consumers use points or cash depending on their willingness-to-pay in cash as
well as in points. Then, we incorporate this consumer choice model into the
seller’s dynamic pricing model, where the revenues from both posted price and
reimbursement for reward sales are embedded in each period.
Sponsored Session
Chair: Narayan Ramasubbu, University of Pittsburgh,
354 Mervis Hall, Pittsburgh, PA, 15228, United States of America,
nramasubbu@katz.pitt.edu
1 - Business Value of the Mobile Enterprise: An Empirical Study of
Mobile Sales Force in Banking
Ajit Sharma, Ross School of Business, 701 Tappan Street,
Ann Arbor, MI, United States of America, asharmaz@umich.edu
3 - An Off-price Retailer with Two Ordering Opportunities
Moutaz Khouja, Professor, UNC Charlotte, 9201 University City
Blvd, Charlotte, NC, 28223, United States of America,
mjkhouja@uncc.edu, Jing Zhou
We develop a model of an off-price retailer who has two procurement
opportunities for next season. The first opportunity occurs after the end of the
current season where she buys excess inventory from retailers and manufacturers
and store them until next season. The second opportunity occurs before the
selling season begins again. The product quantity available in the first opportunity
is limited while the price in the second opportunity is a random variable that
depends on consumer demand.
The press and research on mobility has remained focused on the customer-firm
interface. While there is sufficient evidence of gains from mobile marketing in
better targeting and lift, the benefits of mobile-centric enterprise processes remain
under-studied. In this paper, we empirically assess the reduction in process time
and error rates by shifting from a traditional “sales person in the field-computer in
the office” sales process to a “sales person in the field with a tablet” sales process.
2 - Lost in Cyberspace: An Investigation of Digital Borders, Location
Recognition, and Experience Attribution
Brian Dunn, Assistant Professor, University of Oklahoma, 307
West Brooks Ste. 307D, Norman, OK, 73072, United States of
America, bkdunn@ou.edu, Narayan Ramasubbu, Dennis Galletta,
Paul Lowry
4 - Multi-product Price Promotions with Reference Price Effects
Kevin Li, UC Berkeley, IEOR Department, Berkeley, CA, 94720,
United States of America, kbl4ew@berkeley.edu, Candace Yano
We consider a retailer’s problem of setting prices, including promotional prices,
over a multi-period horizon for multiple products with correlated demands,
considering customer stockpiling and the effect of reference prices on customers’
buying behavior. These factors limit the efficacy of deep discounts and frequent
promotions. We present structural results and numerical examples that provide
insight into the nature of optimal policies and the impact of various parameters.
Do website users know where they are? Given that they may visit multiple sites
in the same session, they may not, which has important implications for the
owners of those sites. However, past research has yet to account for this
possibility. To understand when users recognize where they are online and how
they attribute credit to the sites that are helpful to them, we introduce the
292
■ TB27
TB29
questions about the job. Although the ordeal positively selected candidates, it was
the information in the answers that mattered for match formation. Although the
overall number of matches and speed to fill a vacancy was unchanged, employers
engaged in less recruiting activities and formed higher quality matches.
Multiple Criteria Decision Aiding
2 - Experiments as Instruments: Heterogeneous Position Effects in
Sponsored Search Auctions
Justin Rao, Researcher, Microsoft Research, 641 Avenue of
Americas, New York, NY, 10014, United States of America,
Justin.Rao@microsoft.com, Matthew Goldman
Sponsored Session
Chair: Roman Slowinski, Prof., Poznan University of Technology, Pl.
Marii Sklodowskiej-Curie 5, Poznan, PL, 60-965, Poland,
roman.slowinski@cs.put.poznan.pl
1 - FITradeoff: Flexible and Interactive Tradeoff Elicitation Procedure
Adiel T. DeAlmeida, Professor, Universidade Federal de
Pernambuco, Caixa Postal 7462, Recife, PE, 50630-971, Brazil,
almeidaatd@gmail.com, Adiel De Almeida Filho,
Jonatas Araujo De Almeida, Ana Paula Costa
The Generalized Second Price auction has been shown to achieve an efficient
allocation and favorable revenue properties provided the causal impact of ad
position on user click probabilities is a constant the scaling factor for all ads. We
develop a novel method to re-purpose internal business experimentation at a
major search engine and we strongly reject the conventional multiplicativelyseparable model, instead finding substantial heterogeneity of the causal impact of
position on CTR.
FITradeoff is a Flexible and Interactive Tradeoff elicitation procedure for
multicriteria additive models in MAVT scope. The classical tradeoff procedure is
one of the approaches with strongest theoretical foundation. However, behavioral
studies have shown inconsistences of DM during elicitation. The FITradeoff
reduces DM’s effort in the process, by using partial information, thereby
contributing for reducing inconsistences. It is implemented in a DSS, which is
illustrated by applications.
3 - Optimal Design of Two-sided Market Platforms: An Empirical
Case Study of Ebay
Brent Hickman, Assistant Professor Of Economics, University of
Chicago, 1226 E 59th St, Chicago, IL, 60637, United States of
America, hickmanbr@uchicago.edu, Joern Boehnke,
Aaron Bodoh-creed
2 - An Enhanced “Election Machine” for the Finnish Parliamentary
Elections: Theory and Implementation
Jyrki Wallenius, Professor, Aalto University School of Business,
Runeberginkatu 22-24, Helsinki, Finland,
jyrki.wallenius@aalto.fi, Tommi Pajala, Akram Dehnokhalaji,
Pekka Korhonen, Pekka Malo, Ankur Sinha
We investigate design of platform markets that house many auctions over time.
We combine a unique dataset with a model of bidding where the option value of
re-entering the market creates incentive for buyers to shade bids below private
valuations in the current period. We show the model is identified using the
Bellman equation for a representative bidder. We estimate the model and
investigate the degree to which eBay is able to reduce transaction costs and
approach the efficient allocation.
Web-based questionnaires to match candidates’ and voters’ views play an
important role in Finland. We have collaborated with Helsingin Sanomat, who
runs the most influential of such questionnaires, to enhance and further develop
it. Our algorithm was tested in last April’s Parliamentary Elections. We describe
our algorithm and the feedback.
4 - Stability of Demand Models Across Policy Reforms: An Empirical
Study with Boston Public Schools
Peng Shi, MIT Operations Research Center, 1 Amherst Street,
pengshi@mit.edu, Parag Pathak
3 - Multicriteria and Multiobjective Models for Risk, Reliability and
Maintenance Context
Rodrigo J P Ferreira, Assistant Professor, Universidade Federal de
Pernambuco, Av. Professor Morais Rego, 1235., Recife, PE,
50670-901, Brazil, rodjpf@gmail.com, Adiel T De Almeida,
Cristiano A V Cavalcante, Marcelo H Alencar,
Adiel De Almeida Filho, Thalles V Garcez
In counterfactual analysis using demand modelling, an important but seldom
checked assumption is that the proposed reform does not affect the demand
model. We validate this assumption across a major school choice reform in Boston
in 2014. To control for post-analysis bias, we precommit to forecasts before the
reform. We find that while our prediction of the number of applicants were off,
the logit and mixed-logit demand models we fit were stable before and after the
reform.
The use of multiple criteria and multiobjective models in risk, reliability and
maintenance research has increased in recent years. These models may affect the
strategic results of any organization, as well as, human life and the environment.
In such situations, optimal solutions for one objective function cannot be suitable.
These issues are presented according to the reference Multicriteria and
Multiobjective Models for Risk, Reliability and Maintenance Decision Analysis.
■ TB29
Applications of Analytics II
4 - Constructive Preference Learning in Value-driven Multiple
Criteria Sorting
Roman Slowinski, Prof., Poznan University of Technology, Pl.
Marii Sklodowskiej-Curie 5, Poznan, PL, 60-965, Poland,
roman.slowinski@cs.put.poznan.pl, Milosz Kadzinski,
Krzysztof Ciomek
Sponsor: Analytics
Sponsored Session
Chair: Tarun Mohan Lal, Mayo Clinic, mohanlal.tarun@mayo.edu
1 - Combating Attrition through New Developments in Transaction
Analytics and Customer Dialogue
Gerald Fahner, Analytic Science Senior Director, FICO, 181 Metro
Drive, San Jose, United States of America, geraldfahner@fico.com
We present an interactive preference learning technique for multiple criteria
sorting driven by a set of additive value functions compatible with a rich
preference information acquired from the user. This information may include: (1)
imprecise assignment examples, (2) desired class cardinalities, and (3)
assignment-based pairwise comparisons. The output results are necessary and
possible assignments, and extreme class cardinalities.
”Silent” attrition remains a costly problem requiring fast detection and insight to
create effective retention offers. Our credit card case study shows how ensemble
models instrumented with low-latency transaction features rapidly detect cardlevel and merchant category-level attrition. We explain our models and relate
performance to profitability. We show how to boost persuasiveness of offers by
customer dialogues to learn their preferences. Using a simulation we illustrate the
value of dialogue.
■ TB28
2 - How Bringing Decision Optimization to the Cloud Will
Democratize Optimization
Susara Van Den Heever, IBM France, 1681 Route des Dolines,
France, svdheever@fr.ibm.com, Xavier Ceugniet, Alain Chabrier,
Stéphane Michel
Empirical Market Design
Cluster: Auctions
Invited Session
Even though Decision Optimization has been used effectively across industries for
decades, it remains under-utilized. Complexity and costs are often cited as barriers
to wider adoption. The emergence of cloud computing, as well as the renewed
emphasis on cognitive analytics platforms, breaks down these barriers to bring the
benefits of optimization to a wider audience. We will demonstrate this vision
through a case study involving IBM Decision Optimization on Cloud, and IBM
Watson Analytics.
Chair: Peng Shi, MIT Operations Research Center, 1 Amherst Street,
pengshi@mit.edu
1 - Market Congestion and Application Costs
John Horton, Assistant Professor, NYU Stern School of Business,
44 West Fourth Street, Kaufman Management Center,
John.Horton@stern.nyu.edu, Ramesh Johari, Dana Chandler
We report the results of an experimental intervention that increased the cost of
applying to vacancies in an online labor market by requiring workers to answer
293
TB30
■ TB31
3 - An Approach to Estimating Customer Lifetime Values for
Apartment Tenants
Jian Wang, Vice President, Research & Development, The
Rainmaker Group, 4550 North Point Parkway, Alpharetta, GA,
30022, United States of America, jwang@letitrain.com
Connected Vehicle Analytics
Sponsored Session
Estimating tenant lifetime values is important for apartment revenue
management. We propose a heuristic approach to predicting renewal likelihoods
and estimating tenant lifetime values. We then present empirical results based on
real apartment data.
Chair: Juan Li, Member of Research Staff, Xerox Innovation Group,
800 Phillips Road, 128-27E, Webster, NY, 14580,
United States of America, Juan.Li@xerox.com
1 - A System for Estimating Traffic Congestion Measures in a
Network using GPS Smartphone
Charles Chung, Vp Products, Brisk Synergies, 295 Hagey Blvd,
1st Flr, Waterloo, Canada, charles.chung@brisksynergies.com
■ TB30
Intelligent Agents and Systems
A smartphone app is developed for logging route data. A platform is then built for
mapping traffic congestion using speed indicators average speed and speed
differential at the link level. The results demonstrate the feasibility and huge
potential our data collection system that can be implemented in any city and sets
the growth for real-time applications for connected vehicles.
Contributed Session
Chair: Mohsen Moghaddam, PhD Candidate, Purdue University,
1155 Anthrop Dr., Apt. 9, West Lafayette, IN, 47906,
United States of America, mmoghadd@purdue.edu
1 - Optimizing Physician to Patient Consults using Robot-based
Virtual Systems
Henry Ibekwe, Post Doctoral Researcher, Independent,
Richmond, United States of America, hibekwe1@gmail.com
2 - Online Travel Mode Identification with Smartphones
Qing He, Assistant Professor, SUNY Buffalo, 313 Bell Hall,
Buffalo, NY, 14051, United States of America,
qinghe@buffalo.edu, Xing Su, Hernan Caceres, Hanghang Tong
We propose an online classification algorithm to detect user’s travel mode using
mobile phone sensors. Our application is built on the latest Android smartphone
with multimodality sensors. By applying a hierarchical classification method, we
achieve high accuracy in a binary classification wheelers/non-wheelers travel
mode, and all six travel modes.
The delivery of quality healthcare for chronically ill patients is burdened by the
limited resources available to physicians and healthcare facilities. We propose the
use of virtual-presence autonomous robot systems to optimize the physician to
patient consultation by minimizing the patient wait-time and maximizing the
number of physician consults given limited resources. We formulate a robotpatient interaction model as a stochastic process and solve using discrete-time
dynamic programming.
3 - Locating Heterogeneous Traffic Sensors to Improve Network
Surveillance Benefit
Xuechi Zhang, Graduate Research Assistant, University of
Maryland, 0147C Eng Lab Bld, University of Maryland,
College Park, MD, 20742, United States of America,
zhangxc90@gmail.com, Ali Haghani
2 - A Study on the Influence of Trust and Distrust Ratings in Social
Networks on Cold Start Users
Sanjog Ray, Assistant Professor, Indian Institute of Management
Indore, Rau Pithampur Road, Indore, 453331, India,
sanjogr@iimidr.ac.in
Optimal placement of traffic sensors is significant to improve urban mobility. In
this study, a mathematical optimization model of deploying heterogeneous
sensors (i.e. Bluetooth sensor and loop detector) to large-scale traffic network is
proposed. Maximizing real-time information report reliability and coverage are
chosen as dual objectives. In addition, the effect of real-time GPS-based probe
vehicle data is also considered. A case study in Washington D.C. area is conducted
for demonstration.
This study examines how cold start users get influenced by the trust and distrust
scores of other users in a social network. We examine the users trusted by cold
start users on the basis of critical parameters: number of trust statements, number
of distrust statements, and number of items rated. We base our findings on our
analysis of the real life Epinions dataset. Our analysis has implications for design
of trust aware recommender systems for cold start users.
4 - Inferring Trajectories for Partial Observations
Juan Li, Member of Research Staff, Xerox Innovation Group, 800
Phillips Road, 128-27E, Webster, NY, 14580, United States of
America, Juan.Li@xerox.com, Moshe Lichman, Padhraic Smyth
3 - Analyzing Inventory Policies in Multi-stage Automatic
Manufacturing Systems
Barin Nag, Professor, Towson University, Department of
E-Business & Technology Ma, Towson, MD, 21252, United States
of America, bnag@towson.edu, Dong-qing Yao, Sungchul Hong
The amount of spatial trajectory data is growing fast with the rapid increased
availability of GPS-embedded vehicles. The trajectory data is mixed with high and
low sampling rate with partial observations. In this study, we aim to build
probabilistic models to infer possible traversed route for low sampling rate vehicle
trajectory data.
In a multi-stage manufacturing system each stage fills demand from any
combination of buffer inventory or production. Lowest inventory levels may not
be lowest cost, with contradictions arising from the costs of delays of physical
production, backlogs, breakdowns, and bottlenecks. We study best performance
inventory policies using varied production architectures.
4 - A Modeling Framework of Cyber-Physical Systems
Ashutosh Nayak, Student, Purdue University, 318 N Salisbury St,
Apt. 8, West Lafayette, IN, 47906, United States of America,
nayak2@purdue.edu, Shimon Y. Nof, Seokcheon Lee,
Rodrigo Levalle
■ TB32
Effective modelling of CPS is a big challenge. In this work, we propose a resource
sharing based framework for CPS aimed at maximizing its utility. This framework
represents CPS as a network of tasks and resources characterized by utility
functions and overlapping resource communities. A distributed control approach
backed by utility aggregation function is considered for optimality and stability. Its
implementation is illustrated through two examples: Smart factory and multirobot system.
Sponsor: Analytics
Sponsored Session
Business Analytics in Higher Education Industry
Chair: Roger Gung, Director, Business Analytics & Operations Research,
University of Phoenix, 3137 E Elwood St, Phoenix, AZ, 85034,
United States of America, roger.gung@phoenix.edu
1 - Marketing Mix Optimization
Roger Gung, Director, Business Analytics & Operations Research,
University of Phoenix, 3137 E Elwood St, Phoenix, AZ, 85034,
United States of America, roger.gung@phoenix.edu
5 - Collaborative Networked V-organizations: Design & Integration
Mohsen Moghaddam, PhD Candidate, Purdue University, 1155
Anthrop Dr., Apt. 9, West Lafayette, IN, 47906, United States of
America, mmoghadd@purdue.edu, Shimon Y. Nof
Marketing spend allocation drives the volume of new marketing inquiries (NMI)
and enrollments. Two-stage non-linear regression models were built to formulate
NMI channels with respect to marketing spends which were defined as either
endogenous, exogenous or instrument variables. The optimization model was
formed by aggregating all NMI channels’ regression models into one objective
function. The optimal spend allocation was then derived from the model every
quarter to guide marketing strategies.
Modern distributed, networked, and collaborative organizations of
humans/machines/firms enable systematic integration of distributed resources for
processing dynamic/diverse tasks. We design collaborative networked vOrganizations by integrating physical (location of resources) and virtual
(allocation of tasks) dimensions, for higher service level, stability, and utilization.
A mixed-integer program and a tabu search are developed for modeling and
optimization purposes, respectively.
294
2 - Contact Center Qualifying Transfer Rate Modeling and Analysis
Jie Yu, Operations Research Scientist, University of Phoenix,
3137 E Elwood St, Phoenix, AZ, 85034, United States of America,
jie.yu@phoenix.edu, Roger Gung, Lin Wang
TB34
4 - Strategies for Ebola Containment: A Biological-behavioraloperational Modeling Decision Framework
Eva Lee, Georgia Tech, Atlanta, GA, eva.lee@gatech.edu
This work is joint with CDC. We present a computational decision modeling
framework that integrates an agent-based biological disease spread model, a
dynamic network-based social-behavior model that captures human behavior and
interaction, and a stochastic queueing model that describes treatment
characteristics, day-to-day hospital and homecare processes, and resource usage.
An optimization engine determines the minimum resource needed to contain the
Ebola epidemic in W. Africa.
Impact analysis on transferring marketing inquiries to qualifying leads for
potential enrollments and the performance of contact center agents are crucial to
contact center and enrollment operations. An impact analysis was conducted on
transfer rate with drivers including speed to lead, lead source, time of request as
well as program level. A mixed effect logistic regression model was built to rank
agents’ performance in terms of expected transfer rate with given marketing
inquiries. The model was also employed to evaluate the impact of reducing
contact center and enrollment operating hours.
3 - Enrollment Service Contact Strategy Optimization
Pan Hu, Operations Research Scientist, University of Phoenix,
3137 E Elwood St, Phoenix, AZ, 85034, United States of America,
pan.hu@phoenix.edu, Yun Ouyang, Jie Yu, Lin Wang,
Roger Gung
■ TB34
This project is to study how contact behaviors of enrollment representatives
influence enrollment progression of higher education pursuers. To better serve
the needs of potential students, it is critical to communicate effectively by
bringing up right topics in the best timing. We examined a list of conversation
topics suggested in the internal guideline of University of Phoenix for enrollment
representatives, and identified the best contact strategy using statistical models.
Sponsored Session
Data-driven Modeling and Analysis of Health
Care Systems
Chair: Anil Aswani, UC Berkeley, 4141 Etcheverry Hall, Berkeley, CA,
94720-1777, United States of America, aaswani@berkeley.edu
1 - Constructing Behavioral Models for Personalized Weight Loss
Interventions using Integer Programming
Yonatan Mintz, Graduate Student, UC Berkeley, 1822 Francisco
St., Apt. 10, Berkeley, CA, 94703, United States of America,
ymintz@berkeley.edu, Philip Kaminsky, Yoshimi Fukuoka,
Anil Aswani, Elena Flowers
■ TB33
Joint Session HAS/MSOM-Healthcare: Modeling
Applications for Emergency Departments
In this paper we describe two (a machine learning and a utility maximization)
models for weight loss using clinical trial data. We believe these quantitative
models of behavior change can be used to provide personalized interventions,
improve adherence and lower costs of current weight loss programs. Given the
high prevalence of obesity, these results provide significant insight into more
effective approaches to implement weight loss programs.
Sponsored Session
Chair: Sean Barnes, University of Maryland, 4352 Van Munching Hall,
University of Maryland, College Park, MD, 20742,
United States of America, sbarnes@rhsmith.umd.edu
1 - Review of Queueing Theory Applied to Emergency Departments
with Comparable Simulation Studies
Summer (Xia) Hu, PhD Student, University of Maryland,
Department of Mathematics, College Park, United States of
America, xhu64@umd.edu, Sean Barnes, Bruce Golden
2 - Modeling Treatment Adherence Behavior in the Treatment of
Obstructive Sleep Apnea
Yuncheol Kang, Pennsylvania State University, 236 Leonhard
Building, State College, 16801, United States of America,
kang.yuncheol@gmail.com, Paul Griffin, Vittal Prabhu,
Amy Sawyer
We target patients who suffered from Obstructive Sleep Apnea (OSA) and their
treatment behaviors when using Continuous Positive Airway Pressure (CPAP)
devices. We model underlying dynamics and patterns of patient treatment
behavior using Markov models as a basis for designing effective and economical
intervention. Also we suggest a guideline for designing a cost-effective
intervention to economically treat the patients.
Queueing Theory (QT) is an important tool for Emergency Department (ED)
design and management. By reviewing all papers with ED QT analysis or
applications since 1972, this survey examines the contributions of QT to modeling
EDs and identify its benefits and limitations when compared to discrete-event
simulation (DES) under similar ED operational settings. Our results indicate that
the combination of queueing and DES methods can be a powerful approach to
better ED modelling.
3 - Inverse Optimization with Noisy Data
Auyon Siddiq, UC Berkeley, 4141 Etcheverry Hall, University of
California, Berkeley, Berkeley, 94720, United States of America,
auyon.siddiq@berkeley.edu, Zuo-jun Max Shen, Anil Aswani
2 - Using Simulation to Assess the Impact of an Observation Unit in a
Pediatric Emergency Department
Mark Grum, University of Michigan, 1205 Beal Avenue, Ann
Arbor, MI, 48109, United States of America, mgrum@umich.edu,
Gabriel Zayas-Caban, Michelle Macy, Allison Cator, Amy Cohn
We present an approach for inverse parametric optimization with noisy solution
data for convex forward problems. The proposed method yields well-behaved
estimates that attain risk consistency or parameter estimation consistency under
reasonable conditions. While the formulation is non-convex in general, we
provide an approximation algorithm that yields consistent estimates for a class of
quadratic programs. Numerical results show competitive performance with stateof-the-art techniques.
Observation units (OUs) provide an alternative disposition decision for ED
patients who may benefit from further observation, such as those are not ill
enough to be admitted, but not well enough to be discharged. Patients can be
placed in an OU for monitoring, diagnostic evaluation, and/or treatment prior to
disposition. In this talk, we discuss our approaches (e.g. simulation) for assessing
the impact of an OU in the Pediatric ED at the University of Michigan.
4 - Quantifying the Resilience of Hospital Unit Management under
High Workloads
Mo Zhou, PhD Student, UC-Berkeley, 4470 Etcheverry Hall,
Berkeley, CA, 94709, United States of America,
mzhou@berkeley.edu, Anne Miller, Anil Aswani, Jason Slagle,
Daniel France
3 - Operational Causes of Patients Leaving Before Treatment is
Completed in Emergency Departments
David Anderson, Assistant Professor, Baruch,
davidryberganderson@gmail.com, Bruce Golden, Edward Wasil,
Laura Pimentel, Jon Mark Hirshon
Hospital unit shifts with high admissions/discharges (ADTs) and low nurse-topatient ratios (NPRs) increase mortality. Nurse managers promote unit resilience,
and we quantify this using time series and network analysis of hourly phone calls,
ADTs, and NPRs over 2 years from an Intensive Care Unit. Statistical variable
selection assessed variable dependency, and time-series estimation demonstrated
the validity of phone calls as a resilience measure. Future studies will elucidate
adaptive limits.
Patients leaving before treatment (LBTC) is completed is an indicator of poor
Emergency Department performance. Contrary to previous research, volume is
not the main driver of patients leaving before treatment is complete. First
provider time and lengths of treatment are much more strongly associated with
LBTC rate. We show that operational factors such as treatment time and staffing
decisions play a role in waiting time and, thus, in determining the LBTC rate.
295
TB35
■ TB35
3 - Exploring a Community’s Perception of Resilience and its Effect
on Public Policy
Roberta Russell, Professor, Virginia Tech, 1007 Pamplin Hall,
0235, Blacksburg, VA, 24060, United States of America,
rrussell@vt.edu, Yuhong Li, Michelle Seref
Using Technology to Enhance Guest Experiences and
Performance in Hospitality Management
Resilience has been used in many disciplines to describe the ability of an entity to
withstand the effects of a disaster, to bounce back from a disaster, or to change
and adapt to a new reality post-disaster. This research uses text mining to explore
the use of resilience and related terms in newspaper and other media to describe
disasters and subsequent recovery efforts. Particular views of resilience are
correlated with community actions related to building resilience for the next
disaster.
Cluster: Hospitality, Tourism, and Healthcare
Invited Session
Chair: Alex Susskind, Cornell University, 350 Statler Hall, School of
Hotel Adminstration, Ithaca, NY, 14853, United States of America,
ams76@cornell.edu
1 - Picturing Hotels: Attributes of Hotel Images That Attract
Consumer Attention Online
Stephani K. A. Robson, Senior Lecturer, Cornell University,
255 Statler Hall, Ithaca, NY, 14853, United States of America,
skr4@cornell.edu, Breffni Noone
4 - Improving Resource Pre-positioning to Support Disaster
Relief Operations
Andrew Arnette, Assistant Professor Of Decision Sciences and
Governor Geringer Scholar, University of Wyoming, 1000 E.
University Ave, Laramie, WY, 82071, United States of America,
aarnette@uwyo.edu, Chris Zobel
Images have been shown to be an important element in the online hotel choice
process. This exploratory study uses eye tracking technology to investigate the
attributes of hotel images that attract consumers’ eye fixations during a
naturalistic search for lodging, with implications for hotel marketing strategies.
This research seeks to address a need for improving asset location modeling for
opening overnight shelters in response to natural disasters. Such pre-positioning
is crucial for organizations that provide immediate relief to impacted populations,
and we discuss a mathematical programming approach that improves on previous
attempts to determine more optimal placements for a combination of resources.
2 - The Connection Between Restaurant Performance and
Customer-facing Technology in Restaurants
Alex Susskind, Cornell University, 350 Statler Hall, School of
Hotel Adminstration, Ithaca, NY, 14853, United States of
America, ams76@cornell.edu
■ TB37
In this study, the relationship between customer-facing technology, customer
satisfaction and restaurant performance are examined. The findings suggest that
the use of customer-facing technology in a full-service restaurant experience is
positively connected to customers’ satisfaction with their service experience in the
restaurant, higher average spending in the restaurant and a higher tip percentage
paid to the servers.
Health Care Modeling and Optimization X
Contributed Session
Chair: Songinan Zhao, Student, Kansas State University, 1600 Hillcrest
Dr., Apt 4, Manhattan, KS, 66502, United States of America,
songnian@ksu.edu
1 - Use of Simulation-optimization Technique in Operating
Room Scheduling
Musa Demirtas, Research And Teaching Assistant, Western New
England University, 1215 Wilbraham Road, Springfield, MA,
01119, United States of America, demirtasmusa@gmail.com,
Mohammad Dehghani, Thomas K. Keyser
3 - Does Customer-facing Technology Reduce Service Time
in Restaurants?
Ben Curry, Data Scientist, Ela Carte, Ela Carte Headquarters, San
Francisco, CA, United States of America, bcurry@elacarte.com
Looking at how table turn time and service labor usage was affected by
customers’ use of customer facing technology in restaurants, I found that
customer-facing technology notably reduced table turn time when customers
ordered their meals through the table top device; that figure increased more for
customers who ordered their meals and settled their bills using the table-top
devices.
In hospitals, operating Rooms (ORs) are the most important and costly
departments, and generate a big portion of revenues. This study focuses on
maximization of ORs utilization and minimization of inpatients’ length of stay to
decrease costs and increase patients’ satisfaction. Since the arrival of emergency
patients may disrupt the schedule, we developed a reactive simulationoptimization model to schedules that optimally allocate limited resources to
multiple ORs.
■ TB36
2 - Optimal Delivery and Pickup Planning for Patients with Chronic
Diseases using Drones
Seon Jin Kim, University of Houston, Dept. of Industrial
Engineering, Houston, TX, 77204, United States of America,
sonjin64@gmail.com, Gino Lim, Jaeyoung Cho
Disaster Relief and Humanitarian Logistics
Sponsored Session
Chair: Chris Zobel, Professor, Virginia Tech, Business Info. Technology,
880 W. Campus Dr., Suite 1007, Blacksburg, VA, 24061-0235,
United States of America, czobel@vt.edu
Patients with chronic diseases are required to visit clinics for a routine health
exam. The cost of chronic diseases has been increasing every year, which became
a burden to patients, government, and health insurance companies. We present a
robust optimization model to reduce healthcare costs and improve quality of
healthcare service using drones. The model finds optimal routes of drones to
deliver medicine and pickup necessary samples to analyze patients’ health.
Co-Chair: Andrew Arnette, Assistant Professor Of Decision Sciences
And Governor Geringer Scholar, University of Wyoming, 1000 E.
University Ave, Laramie, WY, 82071, United States of America,
aarnette@uwyo.edu
1 - Management of Blood Supplies During Humanitarian Crises
Cigdem Gonul Kochan, Ohio Northern University, 525 South
Main Street, Ada, OH, 45810, United States of America,
c-kochan@onu.edu, Shailesh Kulkarni, David R. Nowicki
3 - Application of Theory of Constraints in Blood Banking
Harshal Lowalekar, Assistant Professor, Indian Institute of
Management Indore, Rau-Pithampur Road, Indore, MP, 453556,
India, harshal@iimidr.ac.in
We discuss the application of the Theory of Constraints Thinking Processes
(TOCTP) methodology in managing inventory at blood banks. Using the Thinking
Processes approach the root-cause behind the common inventory problems at
blood banks like high shortage and wastage of blood products, high operating
expenses and low revenue levels is identified. A TOC based solution is then
proposed to address the root-cause.
This study presents a combined problem of allocating and routing the available
blood supplies at the central blood bank to a given set of hospitals with uncertain
demand. We develop two multi-product newsvendor (MPNP) –traveling salesman
(TSP) models. We examine and compare the results of both models.
2 - A Markov Decision Process Model for Equitable Distribution of
Supplies under Uncertainty
Lauren Davis, North Carolina A&T State University,
1601 E. Market St., Greensboro, NC, United States of America,
lbdavis@ncat.edu, Sefakor Fianu
4 - Study of Optimal Control Strategies for Visceral Leishmaniasis
Songinan Zhao, Student, Kansas State University, 1600 Hillcrest
Dr., Apt. 4, Manhattan, KS, 66502, United States of America,
songnian@ksu.edu, Chih-hang Wu, Yan Kuang, David Ben-arieh
Visceral Leishmaniasis (VL) is a vector-borne disease which is transmitted by
sandflies and it is the second-largest parasitic killer after Marilia. Mathematical
models were proposed to assist in the control of spread of VL; however,
quantitative conditions for the control of VL transmission are not studied. This
paper develops a general mathematical model for VL disease transmission system,
performs bifurcation analysis to discuss control conditions, and calculates optimal
control strategies.
Food banks are one of many non-profit organizations assisting in the fight against
hunger. Most of the food distributed by the food bank comes from donations
which are received from various sources in uncertain quantities at random points
in time. We present a finite horizon decision-making model that determines the
optimal allocation of supplies to demand locations (charitable agencies) given
stochastic supply.
296
TB40
5 - The Advent of the Intelligent Electronic Health Record
John Glaser, Siemens Healthcare, Malvern, PA,
United States of America, John.Glaser@Cerner.com
approaches – outcome-based management versus behavior-based management.
The relationship characteristics include the length of the relationship, goal conflict
between the buyer and the supplier, and the risk aversion of the two parties.
We’ve made great progress in embedding the Electronic Health Record (EHR) in
our healthcare processes, with use reaching unprecedented rates. Now, we’re
poised to take it the next level with the intelligent EHR. The intelligent EHR will
look very similar to the traditional system – one can still look up patient results
and history and write prescriptions but the application will move past
transactional functions. The intelligent EHR will be characterized by sophisticated
and flexible decision support, rules engines, process monitoring engines,
intelligent displays of important patient data, access to knowledge resources, the
ability to collect data from multiple care settings through a health information
exchange, and tools that enable provider collaboration. The advent of the
intelligent EHR will be necessary if healthcare is to effectively address challenges
such as those generated by payment reform and managing the care of chronically
ill populations.
2 - Nurse Allocation Policy Evaluation and Analysis of Admissions in
an Intensive Care Unit
Osman Aydas, Instructor & PhD Candidate, University of
Wisconsin-Milwaukee, 3202 N Maryland Ave., Milwaukee, WI,
53211, United States of America, otaydas@uwm.edu, Kaan Kuzu,
Anthony Ross
Nurse staffing is a crucial step in providing quality healthcare. Many patient care
units, including Intensive Care Units, have problems in accurately estimating the
number of nurses to use on a daily basis. We evaluate the existing staff allocation
system of an intensive care unit using clinical operational data and develop a
prediction model for estimating the number of admissions to the unit.
3 - Coordinating Contracts for an Express Service Supply Chain
Juzhi Zhang, University of Science & Technology of China, No.96,
JinZhai Road Baohe District, Hefei, 230026, China,
zjuzhi@mail.ustc.edu.cn, Gou Qinglong, Xiaohang Yue
■ TB38
This paper studies the mismatch problem of an express service supply chain, in
which the express company delivers the product from online retailer to
consumers. We show that displaying information on delivery capacity can
decrease consumers’ belief on delay risk and increase the centralized supply
chain’s profit, but it may not be feasible because consumers may not believe it.
We then design some contracts to make the supply chain achieve the profit under
information display.
Queueing Models I
Contributed Session
Chair: Kaan Kuzu, Assistant Professor, University of WisconsinMilwaukee, Lubar School of Business, 3202 N. Maryland Ave.,
Milwaukee, WI, 53221, United States of America, kuzu@uwm.edu
1 - Optimum Staffing of an Outbound Call Center
Doron Feuer, California Polytechnic State University, Pomona,
6493 Joshua St., Oak Park, CA, 91377, United States of America,
doronfeuer@gmail.com, Saar Yaffe, Saeideh Fallah-Fini
4 - Appraisal Viewpoint Disseminate and Evolution Analysis
Online Transaction
Xiening Wang, Associate Professor, DongBei University of Finance
and Economics,China, No 217, JianShan Street, Shahekou Distri,
Dalian, LN, 116025, China, wangxiening@163.com
Through feature analysis and algorithm analysis, this paper researchs the
disseminate and evolution of online transaction appraisal view during
consumption. The conversion rules of consumer view tendency is proposed, by
describing the viewpoint tendency of network users and combining the clustering
properties of the network transactions appraisal viewpoint. It analyzes the ecommerce transactions appraisal of the cellular migration model based on small
world network effects.
This paper couples regression analysis, Markov chain, and queuing theory
techniques and develops a reliable model for optimum staffing of an outbound
call center constrained by irregular inbound lead volume. By tracking trends of
incoming dials throughout the call center our algorithm predicts the optimum
level of staff required to ensure dialing requirements while minimizing wasted
agent time.
2 - Optimal Control of a Queue under Oveflow Probability Constraint
Abdolghani Ebrahimi, Research Assistant, Iowa State University,
225 Washington Ave, Unit 2, Ames, IA, 50010, United States of
America, gebrahimi91@gmail.com, Arka Ghosh
■ TB40
We develop an optimal policy that minimizes the long-run average cost in a
queue system when you have an overflow probability constraint. The queue has a
limited buffer size. As our constraint we define long-run average time that the
buffer is full less than a number. Our objective is to find the optimal policy. We
solve the problem for two cases; one with bounded space and the other with
unbounded space.
40- Room 101, CC
Operations/Sustainability
Contributed Session
Chair: Xin Wang, Carnegie Mellon University, 5000 Forbes Ave,
xinwang1@andrew.cmu.edu
1 - Analysis on the Link Between Corporate Social and Financial
Performance of Korean Business Groups
Donghyup Woo, Doctoral Candidate, State University of New
York at Buffalo, 326 Jacobs Management Center, University at
Buffalo SUNY, Buffalo, NY, 14260, United States of America,
dwoo@buffalo.edu, Nallan Suresh
3 - Analysis of Priority-Based Ticket Queue Data
Kaan Kuzu, Assistant Professor, University of WisconsinMilwaukee, Lubar School of Business, 3202 N. Maryland Ave.,
Milwaukee, WI, 53221, United States of America,
kuzu@uwm.edu, Refik Soyer
We analyze the transactional data from a bank’s priority-based ticket queue
system and estimate the distributions of inter-arrival and service times based on
customer category, day of the week and time intervals in a day. We also predict
the customers’ patience times for each customer category and day of the week,
and develop the corresponding abandonment probabilities using logistic
regression models.
This research identifies contextual relationships between social responsibility and
financial performance in countries transitioning to developed economies. Using
Korean data, it explores how business group affiliation “Chaebol” influences
strategy and social responsiveness. We test the hypothesis that “Chaebol firms”
show strong positive relationship due to institutional pressure and legitimacy
issues.
■ TB39
2 - Do Sustainability Practices Really Help?
Hung-yao Liu, ESC Rennes School of Business, 2 Rue Robert
d’Arbrissel CS 76522, Rennes, France, hungyao.liu@gmail.com,
Rohit Nishant
39-Room 100, CC
Risks Management in Operations/Marketing
Invited Session
We utilize an extensive dataset on the resource consumption of different firms to
investigate the effectiveness of various sustainability practices (i.e., pollution
prevention, product stewardship, clean technology, and sustainability vision) as
conceptualized in Hart’s “sustainability portfolio”, embedded in the Natural
Resource Based View (NRBV). We seek to understand which sustainability
practices help firms achieve resource efficiency.
Chair: Osman Aydas, Instructor & PhD Candidate, University of
Wisconsin-Milwaukee, 3202 N Maryland Ave., Milwaukee WI 53211,
United States of America, otaydas@uwm.edu
1 - Global Supply Chain Social Responsibility: An Agency
Theory Perspective
Xingxing Zu, Morgan State University, 1700 E Cold Spring Lane,
Department of Information Science & Syst, Baltimore, MD,
United States of America, Xingxing.zu@morgan.edu,
Ziping Wang, Yu Xia
This study examines global supply chain social responsibility from a dyad
perspective based on agency theory. We analyze how the characteristics of buyersupplier relationship affect the effectiveness of two different management
297
TB41
3 - Building Temperature Control with User Feedback and
Energy Optimization
John Wen, Professor, Rensselaer Polytechnic Institute, CII 5015,
110 8th St., Troy, NY, 12180, United States of America,
wenj@rpi.edu
4 - A Data-driven Stochastic Model of an Emergency Department
Xiaopei Zhang, Columbia University, 1 Morningside Drive,
Apt. 1710, New York, NY, 10025, United States of America,
xz2363@columbia.edu, Ward Whitt
We explore arrival and length-of-stay (LoS) data from an Israeli Emergency
Department. We fit a time-varying two-class (hospitalized or not) infinite-server
queueing model with nonhomogeneous Poisson arrivals, where the arrival
rate/admission/LoS is periodic over a week/day/day. Departures after long LoS
tend to occur at midnight.
Buildings are occupied by multiple occupants with different comfort preferences
in a shared space. This paper proposes a distributed incentive based strategy to
balance the user comfort feedback and building energy optimization. We establish
the convergence of the proposed algorithm to the optimal temperature set-point
that minimizes the total energy cost and the aggregate discomfort of all occupants.
4 - Green Technology Innovations, Adoption, and Regulation
Xin Wang, Carnegie Mellon University, 5000 Forbes Ave,
xinwang1@andrew.cmu.edu, Alan Scheller-wolf
■ TB42
42-Room 102B, CC
Joint Session MSOM-Health/HAS: Global Health
Delivery
When a government is considering tightening a standard on a pollutant, their
decision often is influenced by the number of firms being able to meet the
tightened standard, because a higher number indicates a more feasible standard.
We study how such regulation may affect a firm’s incentive to develop a new
technology to reduce a pollutant. We find that stricter regulation may discourage
a firm to develop a new technology, but may encourage other firms to adopt the
technology once it is invented.
Sponsored Session
Chair: Jonas Jonasson, Student, London Business School, Regent’s
Park, London, NW1 4SA, United Kingdom, jjonasson@london.edu
1 - Demand vs. Supply-side Investment in Humanitarian Operations
Karthik V. Natarajan, Assistant Professor, University of
Minnesota, 321 19th Avenue South, 3-150, Minneapolis, MN,
United States of America, knataraj@umn.edu,
Jayashankar Swaminathan
■ TB41
41-Room 102A, CC
ED Operations Management
Sponsored Session
Both supply- and demand-side constraints impact program coverage in
humanitarian settings. We first study the problem of identifying the optimal mix
of supply- and demand-side investments faced by a budget-constrained
organization in a centralized setting. We then consider a decentralized setting and
identify the optimal performance-based contract to mobilize demand. In addition,
we also compare the performance of the optimal contract against three contracts
frequently used in practice.
Chair: Vedat Verter, James Mcgill Professor, Desaultels Faculty of
Management, McGill University, 1001 Sherbrooke Street West,
Montreal, QC, H3A 1G5, Canada, vedat.verter@mcgill.ca
1 - Specialist Care in Rural Hospitals: from Emergency Department
Consultation to Ward Discharge
Michael Klein, PhD Candidate, McGill University,
Desautels Faculty of Management, Montreal, Canada,
michael.klein2@mail.mcgill.ca, Vedat Verter, Brian G. Moses,
Hughie F. Fraser
2 - Assessing the Impact of U.S. Food Assistance Delivery Policies
on Child Mortality in Sub-Saharan Africa
Alex Nikulkov, Stanford GSB, 655 Knight Way, Stanford, CA,
94305, United States of America, nikulkov@stanford.edu,
Lawrence Wein
The U.S. is one of the few countries in the world that delivers its food assistance
via transoceanic shipments of commodity-based in-kind food, which is more
costly and less timely than cash-based assistance. Using household survey data,
geospatial data and supply chain modeling, we estimate that child mortality in
sub-Saharan Africa can be reduced by 16.2% if the U.S. switches entirely to cashbased interventions.
Patients often wait for admission to inpatient wards, boarding on stretchers in
hallways. These delays are the key contributor to Emergency Department (ED)
crowding, resulting in adverse effects including higher mortality. We consider the
ED boarding problem from the perspective of specialists. We focus on Internal
Medicine at two hospitals in Nova Scotia, Canada. We propose a stochastic
dynamic programming model to analyze current practice and identify strategies
for improvement.
3 - How Good are Uniform Co-Payments in Increasing
Market Consumption?
Gonzalo Romero, Rotman School of Management, 105 St. George
Street, Toronto, Canada, Gonzalo.Romero@rotman.utoronto.ca,
Retsef Levi, Georgia Perakis
2 - Impact of Coordination and Information Sharing in Urban
Incident Response
Jonathan Helm, Indiana University Bloomington,
1309 E. Tenth Street, Bloomington, IN, United States of America,
helmj@indiana.edu, Alex Mills, Andres Jola-sanchez,
Mohan Tatikonda, Bobby Courtney
We analyze the problem of a central planner allocating co-payment subsidies to
competing heterogeneous firms, under an endogenous market response and a
budget constraint. We present the first worst-case performance guarantees in
maximizing market consumption for the frequently implemented policy of
uniform co-payments. Namely, allocating the same co-payment to each firm is
guaranteed to induce a significant fraction of the optimal market consumption,
even if the firms are highly heterogeneous.
Following a disaster in an urban area, on-scene responders must decide how to
distribute casualties among hospitals. This is typically done without information
(real-time or otherwise) about hospital capacities, ED and inpatient. We study a
new type of organization, called a healthcare coalition, and use real data to study
what types of information this organization should share with responders after a
multiple casualty incident to improve response.
4 - Deployment Guidelines for Community Health Workers in
Sub- Saharan Africa
Jonas Jonasson, Student, London Business School, Regent’s Park,
London, NW1 4SA, United Kingdom, jjonasson@london.edu,
Anne Liu, Sarang Deo, Jérémie Gallien, Carri Chan
3 - Optimal Admission/Discharge Criteria for Patients with Heart
Failure in Observation Units
Sanket Bhat, McGill University, 1001 Sherbrooke Street West,
Room 520, Montreal, QC, H3A 1G5, Canada,
sanket.bhat@mcgill.ca, Beste Kucukyazici, Rick Mah
Community health workers (CHWs) are increasingly important to the delivery of
health care in many African countries. Leveraging an extensive dataset featuring
time, clinical findings and GPS information for CHW visits in Ghana, we develop
a stochastic model describing the health dynamics of a population served by a
time-constrained CHW. This model supports the design of managerial guidelines
for the patient prioritization, catchment area assignment and task profile
definition in a CHW operation.
Although more than 80% of the patients presented to Emergency Departments
with symptoms related to acute decompensated heart failure (ADHF) are
hospitalized, the majority of patients are not in need of an acute intervention
beyond decongestion. These patients could be managed in observation units and
be discharged without hospitalization. We develop a stochastic model that
dynamically assess the risk levels of ADHF patients, and determine criteria to
optimally discharge, observe, or admit them.
298
■ TB43
TB45
2 - The Big Data Newsvendor: Practical Insights from
Machine Learning
Gah-Yi Vahn, Assistant Professor, London Business School,
gvahn@london.edu, Cynthia Rudin
43-Room 103A, CC
Choice Modeling and Assortment Optimization
Sponsored Session
We study the newsvendor problem when one has n observations of p features
related to the demand as well as demand data. Both low- and high-dimensional
data are considered. We propose Machine Learning (ML) and Kernel
Optimization (KO) approaches, and derive tight bounds on their performance. In
a nurse staffing case study we find that the best KO and ML results beat best
practice by 23% and 24% respectively.
Chair: Vineet Goyal, Columbia University IEOR department, 500 West
120th Street, 304 Mudd, New York, NY, 10027, United States of
America, vg2277@columbia.edu
1 - Approximation Algorithms for Dynamic Assortment
Optimization Models
Ali Aouad, Massachusetts Institute of Technology, 77
Massachusetts Avenue, Bldg. E40-149, Cambridge, MA, 02139,
United States of America, aaouad@mit.edu, Danny Segev,
Retsef Levi
3 - Applying Machine Learning to Revenue Management at Groupon
David Simchi-levi, Professor, Massachusetts Institute of
United States of America, dslevi@mit.edu, Alexander Weinstein,
He Wang, Wang Chi Cheung
We propose a new data-driven pricing algorithm for online retailers, which learns
customer demand from online transaction data. Our method first generates
multiple demand functions using a clustering algorithm, and then learns on the
fly which demand function is more likely to be correct. We will also discuss some
field experiment result through collaborating with Groupon, a large daily deal
website.
We study the joint assortment and inventory management problem, where
demand consists in a random sequence of heterogeneous customers. Although
the problem is hard in general, we provide the first polynomial time algorithms
that attain constant approximations, for variants proposed in previous literature
as well as more general choice models. In addition, our algorithms provide
practical means for solving large-scale instances and for incorporating more
realistic contraints.
4 - Demand Forecasting when Customers Consider, Then Choose
America, yliu2@stern.nyu.edu, Srikanth Jagabathula
2 - Capacity Constrained Assortment Optimization under the Markov
Chain Based Choice Model
Chun Ye, Columbia University IEOR department, 500 West 120th
Street, Mudd 315, New York, NY, 10027, United States of
America, cy2214@columbia.edu, Danny Segev, Vineet Goyal,
Antoine Desir
We consider the problem of demand forecasting when customers choose by first
forming a consideration set and then choosing the most preferred product from
the consideration set. The consideration set is sampled from a general model over
subsets. We propose techniques to estimate such models from purchase
transaction data.
We consider a capacity constrained assortment optimization problem under the
Markov Chain based choice model proposed by Blanchet et al. We first show that
even severely-restricted special cases are APX-hard. We then present a constant
factor approximation for the general problem. Our algorithm is based on a “localratio” method that allows us to transform a non-linear revenue function into a
linear function over appropriately modified item prices.
■ TB45
45-Room 103C, CC
3 - Assortment Optimization under a Random Swap Based
Distribution over Permutations Model
Antoine Desir, Columbia University IEOR department, 500 West
120th Street, Mudd 315, New York, NY, 10027, United States of
America, ad2918@columbia.edu, Vineet Goyal, Danny Segev
Revenue Management for Marketing
Sponsored Session
Chair: John Turner, Assistant Professor, University of California, Irvine,
Room SB2 338, The Paul Merage School of Business, Irvine, CA,
92697-3125, United States of America, john.turner@uci.edu
1 - Scheduling of Guaranteed Targeted Display Advertising under
Reach and Frequency Requirements
Ali Hojjat, University of California Irvine, Paul Merage School of
Business, Irvine, CA, 92697, United States of America,
hojjats@uci.edu, John Turner, Suleyman Cetintas, Jian Yang
We consider a special class of distribution over permutations model based on
modeling the consumer preferences by a random number of random swaps from
a small set of fixed preference lists. This model is motivated from practical
applications where preferences of “similar” consumers differ in a small number of
products. We present polynomial time approximation schemes for capacity
constrained assortment optimization problem under the random swap based
distribution over permutation model.
4 - Design of an Optimal Membership Promotion Policy
with Experiments
Spyros Zoumpoulis, Insead, spyros.zoumpoulis@insead.edu,
Duncan Simester, Artem Timoshenko
We propose a novel mechanism for the scheduling of guaranteed targeted
advertising in online media. We consider a new form of contract in which
advertisers specify the number of unique individuals (reach) and the minimum
number of times (frequency) each individual should be exposed. We further
integrate a variety of new features such as desired diversity and pacing of ads over
time or the number of competing brands seen by each individual. We perform
extensive numerical tests on industry data.
Deciding what customer to target with what type of membership promotion is
among the most important decisions that wholesale clubs face. We use the results
of a large-scale membership promotion field experiment involving multiple types
of membership promotions to propose various promotion policies, each relying on
a different algorithm for customer segmentation. We then evaluate the
performance of the proposed policies as implemented in a large-scale field test.
2 - Transaction Attributes and Customer Valuation
Michael Braun, Associate Professor Of Marketing, Southern
Methodist University, 6212 Bishop Blvd., Fincher 309, Dallas, TX,
75275, United States of America, braunm@smu.edu, Eli Stein,
David Schweidel
■ TB44
We propose a model of customer value and marketing ROI that incorporates
transaction-specific attributes and unobserved heterogeneity. From this model,
one can estimate an upper bound on the amount to invest in retaining a
customer. This amount depends on the recency and frequency of past customer
purchases. Using data from a B2B service provider, we estimate the revenue lost
by the firm when it fails to deliver a customer’s requested level of service.
44-Room 103B, CC
Machine Learning in Operations
Sponsored Session
3 - Auctions with Dynamic Costly Information Acquisition
Negin Golrezaei, Auctions With Dynamic Costly Information
Acquisition, University of Southern California, Bridge Memorial
Hall, 3670 Trousdale Parkway, Los Angeles, CA, 90089, United
States of America, golrezae@usc.edu, Hamid Nazerzadeh
Chair: Srikanth Jagabathula, NYU, 44 West Fourth Street, New York,
United States of America, sjagabat@stern.nyu.edu
1 - Prediction vs Prescription in Data-driven Pricing
Nathan Kallus, MIT, 77 Massachusetts Ave., E40-149, Cambridge,
MA, 02139, United States of America, kallus@mit.edu,
Dimitris Bertsimas
We study the mechanism design problem for the seller of an indivisible good in a
setting where buyers can purchase the additional information and refine their
valuations for the good. This is motivated by information structures in online
advertising where advertisers can target users using cookie-matching services. For
this setting, we propose a rich class of dynamic mechanisms, called Sequential
Weighted Second-Price, which encompasses the optimal and the efficient
mechanisms as special cases.
We study the problem of data-driven pricing and show that a naive but common
predictive approach leaves money on the table. We bound missed revenue
relative to the prescriptive optimum, which we show is unidentifiable from data.
We provide conditions for identifiability and appropriate pricing schemes. A new
hypothesis test shows that predictive approaches are practically insufficient while
parametric approaches often suffice but only if they take into account the
problem’s prescriptive nature.
299
TB46
■ TB47
4 - Learning, Revising, and Forgetting Multidimensional Contextual
Features for Online Ad Selection
John Turner, Assistant Professor, University of California, Irvine,
Room SB2 338, The Paul Merage School of Business, Irvine, CA,
92697-3125, United States of America, john.turner@uci.edu,
Tianbing Xu, Amelia Regan, Yaming Yu
47-Room 104B, CC
Supply Chain Social Responsibility
Operations
Sponsored Session
We study how best to match ads to viewers using high-dimensional contextual
features (demographic, browsing behavior) to predict click-through probability.
Using Thompson Sampling in a Bayesian framework, our model learns the
importance of contextual features while adapting/forgetting over time, capturing
changing individuals’ tastes and shifts in the viewing population’s composition.
Chair: Robert Swinney, Associate Professor, Duke University, 100
Fuqua Dr, Durham, NC, 27708, United States of America,
robert.swinney@duke.edu
1 - Supply Chain Social and Environmental Performance:
Measurement, Improvement and Disclosure
Basak Kalkanci, Georgia Institute of Technology,
800 W Peachtree St. NW, Atlanta, GA, United States of America,
Basak.Kalkanci@scheller.gatech.edu, Erica Plambeck
■ TB46
46-Room 104A, CC
Service Models in MSOM
Firms are beginning to measure the social and environmental impacts associated
with their products and (in a few cases) report those impacts to investors and
consumers. Supply chain strategy and structure influence a firm’s costs and
benefits from impact measurement, reduction and disclosure. We evaluate how a
mandate for disclosure affect impacts, firm expected profit, and its valuation by
investors.
Sponsored Session
Chair: Opher Baron, University of Toronto, 105 St George St,
Toronto, ON, Canada, opher.baron@rotman.utoronto.ca
1 - Worker Flexibility Training and Production Decision Rights
Gad Allon, Professor, Kellogg School of Management,
Northwestern University, 2001 Sheridan Road,
Evanston, IL, 60201, United States of America, gallon@kellogg.northwestern.edu, Achal Bassamboo, Evan Barlow
2 - Impact of Supply Chain Transparency on Sustainability under
NGO Scrutiny
Shi Chen, Assistant Professor, University of Washington,
Michael G. Foster School of Business, University of Washington,
Seattle, WA, 98195, United States of America, shichen@uw.edu,
Qinqin Zhang, Yong-Pin Zhou
We explore the interaction between production decision rights and workers’
decisions on training to become flexible resources. Research on flexible resources
is prevalent in the operations management literature. Human resources, however,
are decision makers and have rights to decide on their own training levels. Many
firms, however, have also given workers some production decision rights. We
show how the workers’ training decisions are affected by the identity of the
production decision maker.
We study the use of supply chain transparency as an effective tool to mitigate
supply chain sustainability issues, and in particular, whether the buyer should
reveal her supplier list, knowing that revealed suppliers could face a different
level of NGO scrutiny than the unrevealed ones. We incorporate the strategic
interactions among a buyer, her suppliers, and the independent NGOs.
3 - Responsible Sourcing via Vertical Integration and
Horizontal Sourcing
Adem Orsdemir, Assistant Professor, University of California
Riverside, School of Business Administration, Anderson Hall,
Riverside, CA, 92507, United States of America,
adem.orsdemir@ucr.edu, Bin Hu, Vinayak Deshpande
2 - Revenue Maximization for Cloud Computing Services
Cinar Kilcioglu, Columbia Business School, New York, NY, 10027,
United States of America, ckilcioglu16@gsb.columbia.edu,
Costis Maglaras
We study a stylized model of revenue maximization for cloud computing services,
analyze price data traces from the biggest cloud service provider, Amazon, provide
some possible explanation for price spikes based on intuitive asymptotic analysis
arguments in systems with large scale capacity and large market potential, and
ultimately study the revenue maximization problem faced by the service provider
that operates in an infinite capacity system and in a market with multiple
customer types.
Vertical integration is a viable way to achieve responsible sourcing. In a
competitive setting, we analyze a firm’s integration and responsible sourcing
decisions. We find that demand externality and possibility of supplying the
competitor may fundamentally change firms’ behaviors. Furthermore, high
probability of violation detection may discourage responsible sourcing.
4 - Investing in Supply Chain Transparency for Social Responsibility
Leon Valdes, lvaldes@mit.edu, Tim Kraft, Karen Zheng
3 - Admission and Discharge Decisions in Intensive Care Units
Huiyin Ouyang, UNC Department of Statistics & Operations
Research, 318 Hanes Hall, CB# 3260, Chapel Hill, NC, 275993260, United States of America, ouyang5@live.unc.edu,
Serhan Ziya, Nilay Argon
We study a manufacturer’s decisions when the social responsibility performance
of his supplier cannot be perfectly observed. The manufacturer can invest to
increase the transparency of his supply chain and the performance of his supplier.
An NGO may communicate to consumers the true level of social responsibility,
potentially decreasing profits.
We formulate a MDP model for admission decisions in an ICU where patients’
health conditions change over time according to Markovian probabilities, We find
that the optimal decision can depend on the mix of patients in the ICU and
provide an analytic characterization of the optimal policy. We also identify
conditions under which the optimal policy is state-independent.
■ TB48
4 - Tandem Queues with Reneging – Analysis and Insights
Jianfu Wang, Assistant Professor, Nanyang Business School, NTU,
50 Nanyang Avenue, NTU, Singapore, Singapore,
wangjf@ntu.edu.sg, Opher Baron, Oded Berman,
Hossein Abouee Mehrizi
48-Room 105A, CC
Operational Issues in Agriculture
Sponsored Session
This paper considers tandem queueing systems with reneging. We develop a new
technique to solve two dimensional Markov Chains with non-repeating structure.
Our technique can be applied to additional settings and used to derive different
service level measures. We demonstrate this technique on a two-station tandem
queueing model with reneging, which has been considered analytically
intractable.
Chair: Onur Boyabatli, Assistant Professor of Operations Management,
Singapore Management University, 50 Stamford Road 04-01,
Lee Kong Chian School of Business, Singapore, 178899, Singapore,
oboyabatli@smu.edu.sg
1 - Agricultural Cooperative Pricing of Premium Product
Burak Kazaz, Associate Professor, Syracuse University,
721 University Avenue, Syracuse, NY, 13244, United States of
America, bkazaz@syr.edu, Scott Webster, Nur Ayvaz-cavdaroglu
We consider the problem of price-setting by a cooperative for an agricultural
product with the following characteristics: (1) the open-market price for the
product depends on yield and on quality and (2) the quality of the product is
influenced by farmer investments over the growing season. We identify a simple
pricing scheme that shows potential to improve performance, and we characterize
the drivers and the magnitude of performance improvement.
300
2 - Planning for Product Substitution in Seed Business
Saurabh Bansal, Assistant Professor, Penn State University,
405 Business Building, University Park, PA, 16802,
United States of America, sub32@psu.edu
TB50
4 - Experience and Competition Effects in Penny Auctions
Chris Parker, Pennsylvania State University, 411 Business
Building, University Park, PA, 16802, United States of America,
chris.parker@psu.edu, Pranav Jindal, Tony Kwasnica,
Peter Newberry
We present new analytical results to manage seed substitution in the agribusiness
domain, and discuss results of an empirical case study in collaboration with an
industry partner.
The internet has created many new online retail opportunities. One such model is
the penny auction, an ascending first-price auction where bidders pay a fee to bid
and increase the price by a nominal amount. The winning bidder pays the auction
price and receives the item with all other bidders receiving nothing. We utilize a
detailed dataset from a penny auction company to investigate the effects of
experience and competition on bidding behavior and auction outcomes.
3 - Government Intervention and Crop Diversification in Agricultural
Supply Chains
Duygu Gunaydin Akkaya, Stanford GSB, 655 Knight Way,
Stanford, CA, 94305, United States of America,
duygug@stanford.edu, Kostas Bimpikis, Hau Lee
Agricultural supply chains face immense risks including yield and market price
uncertainty. In order to mitigate these risks, farmers can engage in crop
diversification. Governments also take a role in supporting farmers’ income and
implement various subsidies to alleviate poverty in the farmer population. We
study how interventions and diversification practices impact the supply chain in
the presence of random yield and endogenous market price.
■ TB50
50-Room 106A, CC
Value Chain Innovations in Developing Economies
Sponsored Session
4 - Corn or Soybean: Dynamic Farmland Allocation under Uncertainty
Onur Boyabatli, Assistant Professor of Operations Management,
Singapore Management University, 50 Stamford Road 04-01, Lee
Kong Chian School of Business, Singapore, 178899, Singapore,
oboyabatli@smu.edu.sg, Javad Nasiry, Yangfang Zhou
Chair: Saibal Ray, Professor, McGill University, 1001 Sherbrooke Street
West, Montreal, Canada, saibal.ray@mcgill.ca
Co-Chair: Fei Qin, Post-doc Research Fellow, McGill University,
Desautels Faculty of Management, Montreal, QC, Canada
1 - Milking the Quality Test: Improving the Milk Supply Chain under
Competing Collection Intermediaries
Liying Mu, Assistant Professor, University of Delaware,
20 Orchard Rd, Newark, 19716, United States of America,
muliying@udel.edu, Milind Dawande, Xianjun Geng,
Vijay Mookerjee
This paper studies the farmland allocation decision of a farmer between two crops
in a multi-period framework. In each period, the farmer chooses the allocation in
the presence of revenue uncertainty, and crop rotation benefits across periods. We
characterize the optimal policy and investigate the impact of revenue uncertainty.
We propose a heuristic allocation policy which is near-optimal.
■ TB49
We examine the quality issues of milk — via deliberate adulteration by milk
farmers — acquired by competing collection intermediaries in developing
countries. Interestingly, some intuitive interventions such as providing collection
stations with better infrastructure (e.g., refrigerators) or subsidizing testing costs
could hurt the quality of milk in the presence of competition. The goal of this
study is to provide recommendations that address the quality problem with
minimal testing.
49-Room 105B, CC
Retail Operations
Sponsored Session
Chair: Chris Parker, Pennsylvania State University, 411 Business
chris.parker@psu.edu
1 - Supply Chain Structure and Multimarket Competition
O. Cem Ozturk, Assistant Professor Of Marketing, Georgia
Institute of Technology, 800 West Peachtree St. NW., Atlanta, GA,
30308, United States of America,
cem.ozturk@scheller.gatech.edu, Necati Tereyagoglu
2 - Low Cost Cataract Surgery in India: What Can Western Health
Systems Learn from it?
Harish Krishnan, Sauder School of Business, University of British
Columbia, Vancouver, Canada, Harish.Krishnan@sauder.ubc.ca
The Aravind Eye Care System (AECS) in India is known for its low-cost business
model. However, few studies have done a detailed analysis of the cost structure of
a cataract surgery at the AECS. This talk will present cost data from AECS and
compare it to similar data at an eye hospital in Canada. The goal is to identify root
causes of the cost differences between AECS and western health systems. The
barriers to implementing AECS’ innovations in the west will also be discussed.
We study the role of supply chain structure in determining competitive intensity
when manufacturers and retailers encounter in multiple markets. Our theoretical
model shows how the differences in supply network overlap across multiple
markets lead to higher retail prices. Using an extensive scanner data set, we find
empirical support for the analytical results. These findings show the importance of
supply chain structure in assessing multimarket competition among firms.
3 - Multi-treatment Inventory Allocation in Humanitarian Health
Settings under Funding Constraints
Jayashankar Swaminathan, UNC-Chapel Hill, 300 Kenan Drive,
Chapel Hill, NC, 27599, United States of America,
Jay_Swaminathan@kenan-flagler.unc.edu, Karthik V. Natarajan
2 - Value of Downward Substitution under Stochastic Prices
Fehmi Tanrisever, Bilkent University, Bilkent, Ankara, Turkey,
tanrisever@bilkent.edu.tr, Zumbul Atan, Junchi Tan
We study the problem of allocating inventory procured using donor funding to
patients in different health states over a finite horizon with the objective of
minimizing the number of disease-adjusted life periods lost. The optimal policy is
state-dependent and hence, we develop two heuristics for the allocation problem.
We also provide analytical results and computational insights regarding how the
funding level and funding timing impact program performance.
Downward substitution as a form of operational flexibility has received significant
academic attention. The literature on downward substitution follows the main
stream inventory literature and assumes uncertain demand and/or yield and
explores the value of substitution flexibility. They, however, assume fixed prices
which may distort the analysis in many industries where prices may fluctuate. In
this paper, we explore the effect of price uncertainty on the value of downward
substitution.
4 - Micro-entrepreneurship in Agri-food Supply Chains in
Developing Economies
Fei Qin, McGill University, Desautels Faculty of Management,
Montreal, QC, H3A1G5, Canada, fei.qin@mcgill.ca,
Mehmet Gumus, Saibal Ray
3 - Supply Chain Contracts that Prevent Information Leakage
Yiwei Chen, Assistant Professor, Renmin University of China, NO.
59 Zhongguancun Street, Beijing, China, chenyiwei@rbs.org.cn,
Ozalp Ozer
Motivated by Veggie-Kart direct farm-to-food initiative for marginal farmers and
retailers in developing economies, we examine the impact of supply chain
innovations involving micro-entrepreneurs at both upstream and downstream
stages, which compete with the traditional spot-market based channel in presence
of supply uncertainty.
We study a supply chain with one supplier and two competing retailers
(incumbent and entrant). The incumbent has better but imprecise private
forecast. We explore general conditions that a wide range of contracts need to
satisfy to prevent the supplier from leaking the incumbent’s private forecast to the
entrant. We define two groups of contracts based on how the supplier and
retailers share inventory risks. We find only these two groups of contracts may
avoid information leakage.
301
TB51
■ TB51
2 - Automatic Feedback Control for Shunt Drainage in
Hydrocephalus Patients
Kalyan Raman, Professor, Northwestern University,
Medill School, Evanston, IL, United States of America,
kalyraman@gmail.com, Vijay Viswanathan
51-Room 106B, CC
Procurement Mechanisms
Sponsored Session
Excessive intracranial pressure (ICP) resulting from insufficient drainage of
cerebrospinal fluid (CSF) leads to a neurological disorder called hydrocephalus,
which is treated by implanting shunts to reduce ICP by draining excess CSF. We
use non-linear control theory to develop a mathematical algorithm for a regulator
to achieve shunt action that is significantly more sophisticated than that of a
switch.
Chair: Tharanga Rajapakshe, Assistant professor, University of Florida,
W. University Ave, Gainesville, FL, 32611, United States of America,
tharanga@ufl.edu
1 - Distressed Selling by Farmers: Model, Analysis, and Use
in Policy-Making
Shivam Gupta, PhD Candidate, UT Dallas, NJ School of
Management, 800 W. Campbell Rd., Richardson, TX, 75080,
United States of America, sxg104920@utdallas.edu,
Ashutosh Sarkar, Ganesh Janakiraman, Milind Dawande
3 - Multiattribute Pricing
Thomas Weber, Associate Professor, EPFL, CDM-ODY 3.01,
Station 5, Lausanne, VD, 1015, Switzerland,
thomas.weber@epfl.ch
We provide a technique for constructing second-best multiattribute screening
contracts in a general setting with one-dimensional types based on necessary
optimality conditions. Our approach allows for type-dependent participation
constraints and arbitrary risk profiles. As an example we discuss optimal
insurance contracts.
The surprising practice of distressed selling, where farmers sell produce to outside
agents at prices much lower than the government’s guaranteed price, is common
in developing countries. We build a tractable stochastic DP model that captures
the ground realities – limited and uncertain procurement capacity, high holding
costs, and lack of affordable credit – that lead to distressed sales. Using real
procurement data, we establish the accuracy of our model and develop useful
policy suggestions.
4 - Dynamic Incentives in Sales Force Compensation
Olivier Rubel, UC Davis, Graduate School of Management,
One Shields Avenue, Davis, CA, United States of America,
orubel@ucdavis.edu, Ashutosh Prasad
2 - Coordinating Procurement Decisions in Multi-division Firms
Fang Fang, Ph. D. Candidate, University of Miami, 5250
University Drive, Coral Gables, FL, 33124, United States of
America, f.fang@umiami.edu, Hari Natarajan
We propose dynamic principal-agent model to investigate how to incentivize sales
people when current selling efforts and carryover sales drive present sales. We
show that the carryover effect increases not only expected sales, but also sales
uncertainty. We then find that the manager incentivizes the high risk-aversion
salesperson with a concave compensation and the low risk-aversion salesperson
with a convex compensation.
Central procurement organizations (CPO) of large firms must coordinate firmwide procurement to leverage volume discounts from suppliers. Facing such a
procurement coordination problem, we examine how a CPO can design internal
prices to maximize firm-wide cost savings. Our analysis of commonly-used
internal pricing rules shows interesting impacts on vendor selection, divisional
participation, and gain allocation.
■ TB53
3 - Does Quality Knowledge Spillover at Shared Suppliers? –
An Empirical Investigation
Suresh Muthulingam, Assistant Professor Of Supply Chain
Management, SMEAL College of Business, The Pennsylvania
State University, 460 Business Building, State College, PA, 16802,
United States of America, sxm84@psu.edu, Anupam Agrawal
53-Room 107B, CC
Behavioral Issues in the OM / Marketing Interface
Sponsored Session
Chair: Ozalp Ozer, The University of Texas at Dallas, 800 West Campbell
Road, Richardson, TX, United States of America, oozer@utdallas.edu
We study the spillover of quality knowledge across supply chains. We observe the
quality performance of 191 suppliers who use the same facilities to manufacture
similar products for two distinct businesses. We find that quality knowledge spills
over under three conditions: (i) When quality efforts focus on organizational
members; (ii) When quality efforts focus on output activities of suppliers; and (iii)
When quality knowledge is developed at suppliers with low complexity.
4 - Contracting Between a Blood Bank and Hospitals
Anand Paul, University of Florida, 351 Stuzin Hall,
Gainesville, FL, United States of America, paulaa@ufl.edu,
Tharanga Rajapakshe
Co-Chair: Upender Subramanian, United States of America,
upender@utdallas.edu
1 - Pricing Cause Marketing Products in the Presence of
Social Comparison
Paola Mallucci, Assistant Professor of Marketing, University of
Wisconsin at Madison, 4261 Grainger Hall, 975 University Ave,
Madison, WI, 53706, United States of America,
pmallucci@bus.wisc.edu, Tony Haitao Cui, George John
The supply of blood at a regional blood bank (RBB) is uncertain and often
insufficient to satisfy the total demand for it. The RBB typically does not observe
the demand at each hospital before determining the allocation policy. Inefficient
allocation leads to shortages at hospitals which necessitates reallocation of blood
and significant blood outdating cost. We make an analytical study of socially
optimal contracting decisions of an RBB serving multiple hospitals.
The broad takeaway from the literature on cause marketing campaigns, where
firms donate to charities with purchase, is that they generally work well, because
of ``warm glow’’. We conjecture that far from creating only positive feelings, such
firm donations can create discomfort by encouraging social comparison. We find
that firms can find it profitable to exploit such discomfort even if it decreases
consumers utility. Results apply in both monopoly and competition.
■ TB52
2 - Pricing and Quality Perception: Theory and Experiment
Karen Zheng, MIT, 77 Massachusetts Avenue, Cambridge, MA,
02139, United States of America, yanchong@mit.edu, Rim Hariss,
Georgia Perakis, Wichinpong Sinchaisri
52-Room 107A, CC
We study how a constant pricing strategy versus a markdown strategy may
induce different perceptions of quality among consumers, and how a firm should
take these quality perceptions into account when optimizing its pricing policy for
competitive products. We empirically elicit the relationship between consumers’
perceived quality and prices under either pricing strategy, and incorporate these
relationships into our consumer model to analyze the firm’s optimal pricing
policy.
Marketing and Optimal Control
Sponsored Session
Chair: Olivier Rubel, UC Davis, Graduate School of Management, One
Shields Avenue, Davis, United States of America, orubel@ucdavis.edu
1 - Optimal Learning to Select the Best Alternative
Tony Ke, Assistant Professor, Marketing Department, MIT Sloan
School of Management, 100 Main Street, E62-535, Cambridge,
MA, 02142, United States of America, kete@mit.edu,
Miguel Villas-boas
3 - Conflict of Interest and Market Structure in Multiplayer Games
Sung Ham, Assistant Professor of Marketing, George Washington
University, 2201 G St. NW, Washington, DC, 20052, United States
of America, sungham@gwu.edu, Jiabin Wu, Noah Lim
When a firm serves customers who compete with one another, a conflict of
interest may arise. We develop a multi-player game where firms serve competing
customers, and examine how the market structure faced by the firms impacts the
extent to which conflicts of interest affect behavior. We test our theory using an
incentive-aligned experiment and find that the decisions are consistent with the
model predictions.
A decision maker is deciding among several alternatives with uncertain payoffs
and an outside option with known payoff. Before making a choice, he can
purchase informative signals on each alternative. We solve for the decision
maker’s optimal learning as well as stopping problem, and discuss the
implications.
302
4 - The Behavioral Cost of Quality Nonconformance:
Risk-averse and Experience-sampling Customers
Jordan Tong, Assistant Professor, University of Wisconsin at
Madison, WI, United States of America, jordan.tong@wisc.edu,
Greg Decroix
TB57
4 - Two Different Approaches to Stochastic DEA
Ole Olesen, Professor, University of Southern Denmark,
Campusvej 55, Odense, 5230, Denmark, ole@sam.sdu.dk,
Niels Chr. Petersen
Focus is on different views on extending DEA to a stochastic setting. The
management science framework does not focus to model performance using a
statistical model based on a specific Data Generating Process (DGP). Some
stochastic DEA models focus on replacing the observed input output observations
with DMU specific distributions. The statistical framework insists on an axiomatic
approach to a statistical model, including a specification of a DGP. We illustrate
these differences.
Why do customers purchase less when quality is inconsistent? A common
explanation is that customers have risk-averse preferences: they inherently prefer
less uncertainty. Another explanation, however, is that tendencies towards lowvariance alternatives are due to a learning process from experience. We show that
optimal pricing and promotion decisions can differ significantly depending on
which explanation is modeled, thereby illuminating the costs of nonconformance
and how to mitigate them.
■ TB56
■ TB54
56-Room 109A, CC
54-Room 108A, CC
Multiple Stakeholders in NPD
Approximations of Queueing Performance for
Rapid Systems Design
Invited Session
Cluster: Tutorials
Invited Session
Chair: Niyazi Taneri, SUTD, 8 Somapah Rd, Singapore, Singapore,
niyazitaneri@sutd.edu.sg
1 - The Role of Decision Rights in Collaborative
Development Initiatives
Nektarios Oraiopoulos, Cambridge Judge Business School,
University of Cambridge, Trumpington St., Cambridge, United
Kingdom, n.oraiopoulos@jbs.cam.ac.uk, Vishal Agrawal
Chair: Ton Dieker, Columbia University, 500 W 120 St, New York, NY,
United States of America, ton.dieker@ieor.columbia.edu
1 - Tutorial: Approximations of Queueing Performance for Rapid
Systems Design
Ton Dieker, Columbia University, 500 W 120 St, New York, NY,
United States of America, ton.dieker@ieor.columbia.edu,
Steve Hackman
In this paper, we study initiatives for co-development of new products and
technologies. In such settings, it may be difficult a priori to specify contracts
contingent on the outcome. Therefore, we investigate the efficacy of different
contractual structures, which instead specify the decision-making process.
Recent advances in queueing analysis have yielded tractable approximations of
performance metrics that can be used to quickly explore initial designs, to reduce
computational burdens associated with simulation, or even to eliminate the need
for simulation altogether. This TutORial takes you on an accessible tour of these
recent methods, shows you how to apply them using numerical examples drawn
from real applications, and discusses implementation challenges and potential
opportunities.
2 - Structuring New Product Development Partnerships
Niyazi Taneri, SUTD, 8 Somapah Rd, Singapore, Singapore,
niyazitaneri@sutd.edu.sg, Arnoud De Meyer
New product development partnerships involve a high degree or risk, information
and incentive problems across various stakeholders. Partners structure their
alliances to address such concerns. We identify factors that affect the structure of
the partnership and the performance of the partnership.
■ TB55
3 - The Impact of Continuous Product Development and Customer
Feedback on Mobile App Performance
Nilam Kaushik, University College London, University College
London, London, United Kingdom, nilam.kaushik.13@ucl.ac.uk,
Bilal Gokpinar
55-Room 108B, CC
Stochastic Methods in Efficiency Analysis
Cluster: Data Envelopment Analysis
Invited Session
Mobile application development differs from traditional product development
owing to low barriers of entry, the ability to provide continuous software updates,
and ease of access to customer feedback. Using a dataset from the App Store, and
drawing from a combination of text mining techniques and econometric methods,
we investigate the impact of incorporating customer feedback on mobile app
performance.
Chair: Ole Olesen, Professor, University of Southern Denmark,
Campusvej 55, Odense, 5230, Denmark, ole@sam.sdu.dk
1 - Estimating Production Functions and Frontiers using
Stochastic DEA
John Ruggiero, Professor, University of Dayton, Dayton, OH,
United States of America, jruggiero1@udayton.edu
■ TB57
In this paper, we present two methods to estimate production functions and
frontiers (deterministic and stochastic). We constrain the technology using the
Afriat conditions and consider minimizing the sum of absolute and/or squared
errors. We extend this method using locally weighted least squares in the spirit of
loess (local regression.)
57-Room 109B, CC
Assorted Topics in Renewable Energy
Climate Change)
Sponsored Session
2 - Endogeneity in Stochastic Frontier Models
Artem Prokhorov, U Sydney, CIREQ, St. Petersburg State U,
Business School, Sydney, NS, 2006, Australia,
artem.b.prokhorov@gmail.com, Peter Schmidt, Christine Amsler
Chair: Anthony Papavasiliou, Université Catholique de Louvain,
Voie du Roman Pays 34, Louvain la Neuve, Ou, 1348, Belgium,
tpapva@hotmail.com
1 - A Controlled Approximation Scheme for Managing Hydroelectric
Generation with Multiple Reservoirs
Bernard Lamond, Professor, Universite Laval, Dep. Operations &
Systemes de Decision, 2325, Rue de la Terrasse #2620, Quebec,
QC, G1V 0A6, Canada, Bernard.Lamond@fsa.ulaval.ca,
Pascal Lang, Pascal Cote, Luckny Zephyr
Stochastic frontier models are typically estimated by MLE or corrected OLS. The
consistency of either estimator depends on exogeneity of the explanatory
variables (inputs, in the production frontier setting). We will investigate the case
that one or more of the inputs is endogenous, in the simultaneous equation sense
of endogeneity. We will consider modifications of standard procedures under
endogeneity for the stochastic frontier setting.
3 - Shape Constrained Kernel Weighted Least Squares for the
Estimation of Production Functions
Andrew Johnson, Texas A&M, College Station, TX,
United States of America, ajohnson@tamu.edu, Daisuke Yagi
We present an approach for adaptive approximation of the value function in
stochastic dynamic programming. We use a simplicial partition of the state space
to construct a nonseparable piecewise affine approximation which is refined
iteratively using lower and upper bounds on the value function. The proposed
scheme is experimented numerically in the context of hydroelectric production
across multiple reservoirs and power plants.
This paper proposes a unifying model and estimator we call Shape Constrained
Kernel-weighted Least Squares (SCKLS). We show the relationship between the
SCKLS estimator and both the Convex Nonparametric Least Squares (CNLS) and
Du’s estimators. Specifically, the SCKLS estimator converges to the CNLS
estimator as the bandwidth goes to zero. We compare the performance of the
three estimators (SCKLS, CNLS, and Du’s estimator) via Monte Carlo simulations.
303
TB58
2 - Demand-side Power Procurement with Renewable Generation
and Energy Storage
Soongeol Kwon, Texas A&M University, 3131 TAMU,
College Station, TX, 77843, United States of America,
soongeol@tamu.edu, Natarajan Gautam
3 - Economics of High-temperature Reactors for Industrial
Cogeneration: A Utility’s Perspective
Reinhard Madlener, RMadlener@eonerc.rwth-aachen.de, Jona
Hampe
This paper studies the economic potential of using HTRs for cogeneration of
industrial process heat and electricity. We find that a reference case HTR can
deliver cost-competitive process heat (at 200 ∞C), thus rendering the chemical
and pulp & paper industries potential candidates. We use real options analysis to
deal with uncertainty and the managerial flexibilities of the project. We also
propose a model to calculate the option of switching between two different
operation modes.
We consider operational decisions to satisfy power demand while minimizing
purchase cost over time-varying electricity prices. In our scenario, consumers use
renewable sources to serve power demand and operate energy storage. We
propose a two-stage stochastic optimization problem to control purchase,
consumption and operations based on day-ahead and real-time procurement
while responding to variability and uncertainty in power demand, renewable
sources and electricity prices.
4 - Selecting the Optimum Nuclear Fuel Cycle Including
Quasi-rational Opinions and Public Perception
Sama Bilbao Y Leon, Associate Professor, Virginia Commonwealth
University, 401 W Main St, Richmond, VA, 23284, United States
of America, sbilbao@vcu.edu,
John Swanson, Ishoc Salaam, Jonathan Hill
3 - Wind Speed Forecasting for Wind Parks: A Sequential
Modeling Approach
Vignesh Subramanian, Dept. of Industrial and Management
Systems Engineering, University of South Florida, ENB118,
Tampa, FL, United States of America, vigneshs@mail.usf.edu,
Tapas K. Das
Although much work has been done to address the technological challenges
associated with the management and ultimate disposal of used nuclear fuel, less
attention has been given to public perception and acceptance of the selected fuel
cycle. This work presents current progress in a decision making model based on
Multi-Attribute Utility Theory that contains the fundamental objectives for both
technical and non-technical factors.
Inherently intermittent nature of wind energy makes it essential to accurately
predict wind speed for reliable operation of power systems comprising wind
generation. We propose a two-stage model. Stage I uses SVM to classify wind
speed into three clusters: zero power, rated power, and continuous power. Stage II
employs a Bayesian additive regression kernel (BARK) method to the continuous
power cluster to estimate the wind speed. The model is tested on numerical
weather prediction (NWP) data.
4 - Capacity Remuneration in the Belgian Electricity Market
Anthony Papavasiliou, Université Catholique de Louvain,
Voie du Roman Pays 34, Louvain la Neuve, Ou, 1348, Belgium,
tpapva@hotmail.com
■ TB59
Belgium experienced a serious shortage in capacity recently due to the unplanned
outage of nuclear capacity. This has motivated an investigation of Belgian capacity
remuneration mechanisms. In this presentation we compare the existing fixed
reserve requirements mechanism with the introduction of operating reserve
demand curves in reserve auctions.
Sponsored Session
■ TB58
This paper attempts to determine the optimal timing and location of fuel
treatments and timber harvests for a multi-stand landscape, accounting for the
spatial interactions that drive the fire behavior. An optimization method known
as value iteration is used to solve the dynamic program. Outcomes for multiple
land ownership configurations are explored.
59-Room 110B, CC
Fire Management 2: Landscape & Modeling
Chair: Hailey Buckingham, hailey.buckingham@oregonstate.edu
1 - Timber Harvest and Fuel Treatment Decisions with Fire Risk
Chris Lauer, cjlauer@gmail.com, Claire Montgomery
58-Room 110A, CC
Topics in Nuclear Energy
Climate Change)
Sponsored Session
2 - Integrating Wildfire Risk and Spread in a Cellular Forest
Harvesting Model
Marc McDill, Pennsylvania State University, University Park, PA,
United States of America, mmcdill@psu.edu, Susete Marques,
José Borges
Chair: Alexandra Newman, Professor, Colorado School of Mines,
Mechanical Engineering, Golden, CO, 80401, United States of America,
anewman@mines.edu
1 - Optimizing the Placement of Radioactive Isotope Measurement
Devices in a Nuclear Fuel Cycle
Ben Johnson, PhD Student, Colorado School of Mines,
Golden, CO, 80401, United States of America,
bebjohns@mymail.mines.edu, Alexandra Newman, Jeffrey King
We present a stochastic, cellular multi-objective forest harvest scheduling model
incorporating a mechanistic model of fire risk probability based on the state of a
cell and the probability of fire in neighboring cells. The model illustrates a
potential approach to integrate management activities including fuel treatments
and harvesting to address multiple objectives.
■ TB60
The purpose of nuclear safeguards is to prevent proliferation of radioactive
material. Enhancing methods to detect potential proliferation will help reduce the
increasing threat of malicious entities successfully obtaining nuclear material. We
create a mixed integer program to determine how many, where, and which types
of radioactive isotope measurement devices should be used in the nuclear fuel
cycle to minimize the weighted sum of Type I and Type II measurement errors.
60-Room 111A, CC
Education II
Contributed Session
Chair: Omar Ben-ayed, Professor of Management, Qatar University,
Al Jameaa Street, P.O. Box 2713, Doha, 2713, Qatar,
omar.benayed@qu.edu.qa
1 - Challenges of Imbedding a Built-In OR/MS Paradigm among
Engineering Graduates
Thong Goh, Professor, National University of Singapore,
1 Engineering Drive 2, Singapore, 117576, Singapore,
tng@nus.edu.sg
2 - Modeling Societal Disruption from Nuclear Accidents to Inform
Regulatory Decision-making
Vicki Bier, Professor, University of Wisconsin - Madison, 1513
University Avenue, 3270A Mechanical Engineering Building,
vicki.bier@wisc.edu, Michael Corradini, Caleb Roh, Shuji Liu,
Robert Youngblood
Nuclear regulation in the U.S. focuses on preventing radiation-related fatalities.
However, recent experience shows that societal disruption from relocation can be
considerable, arguably more significant than radiation-induced health effects. We
have evaluated the population relocation that could occur after severe reactor
accidents as a proxy for societal disruption, and argue that regulatory guidance
should constrain societal disruption as well as radiation exposure.
There are many OR/MS courses at various levels that undergraduates can take.
However, many non-OR/MS disciplines may offer such courses as credit
accumulators, with both teachers and students having them as a subject of
learning and not something permanent that lasts into the students’ subsequent
working lives. This presentation discusses this issue with particular reference to
engineering undergraduate education, and explore ways to alleviate the situation.
304
TB62
2 - Measuring Interculture Competence Among Business, Education
and Social Work Students
Amarpreet Kohli, Assistant Professor, Farmingdale State College,
SUNY, Farmingdale, NY, United States of America,
kohlias@farmingdale.edu, Hermeet Kohli, Cheng Peng
4 - Inducing Environmental Disclosures: A Dynamic Mechanism
Design Approach
Shouqiang Wang, Assistant Professor, Clemson University, 131D
Sirrine Hall, Clemson, SC, 29672, United States of America,
shouqiw@clemson.edu, Peng Sun, Francis De Véricourt
Purpose of this interdisciplinary research was to measure the level of
multicultural awareness, sensitivity to, and understanding of difference, and
intercultural competence when working in diverse environments in the
undergraduate and graduate Business, Education and Human Development, and
Social Work students using the UDO (MGUD-S Survey). Convenient purposive
sampling was utilized to invite students enrolled in these three schools to
participate in web based descriptive survey research.
This paper studies the design of voluntary disclosure regulations that jointly uses
inspections and monetary rewards. We formulate this problem in a dynamic
mechanism design framework with state verification and obtain complete
analytical solution.
■ TB62
3 - Redesigning Qatar University Class Meeting Pattern to
Improve Performance
Omar Ben-ayed, Professor Of Management, Qatar University,
Al Jameaa Street, P.O. Box 2713, Doha, 2713, Qatar,
omar.benayed@qu.edu.qa, Heba Younis, Hend Hammad
62-Room 112A, CC
Aviation Applications
Contributed Session
This study examines the existing class meeting pattern at Qatar University based
on the strategic plan of the University in addition to the perception of students
and faculty members. The study shows that there is a need for a new class
meeting pater with two additional non-teaching half-days. A capacity analysis
proves the feasibility of such a pattern. Accordingly, alternative class meeting
patterns are proposed and one is selected based on technical, academic and
cultural perspectives.
Chair: Jinkun Lee, The Pennsylvania State University, 236 Leonhard,
University Park, United States of America, jinkunlee@psu.edu
1 - An Airspace Sectorization Approach Based on Spectral
Clustering and NSGA-II
Bang An, Tsinghua University, Room 616, Main Building, Beijing,
China, ab13@mails.tsinghua.edu.cn, Peng Cheng, Xiang Zou
■ TB61
We propose an airspace sectorization approach based on spectral clustering and
NSGA-II. With the method embedded in the constrained NSGA-II, all of the
critical constraints can be easily handled. Besides, an initial sectorization method
based on spectral clustering is proposed to generate the first generation of NSGAII. We test our method on the high-altitude airspace controlled by Beijing Area
Control Center. The results show that our method can obtain better solutions.
61-Room 111B, CC
Economics of Reverse Logistics and
Sustainable Operations
2 - A Preemptive Scheduling Model with Overtime Allocation for
Minimum Weighted Tardiness
Fernando Jaramillo, University of Miami, 1251 Memorial Drive,
Department of Industrial Engineering, Coral Gables, FL, 33146,
United States of America, f.jaramillo2@umiami.edu, Busra Keles,
Murat Erkoc
Sponsored Session
Chair: Shouqiang Wang, Assistant Professor, Clemson University,
131D Sirrine Hall, Clemson, SC, 29672, United States of America,
shouqiw@clemson.edu
1 - Versioning, Trade-ins and Refurbishing: An Integrative Analysis
Avinash Geda, University of Florida, 361B Stuzin Hall,
Gainesville, FL, 32611, United States of America,
avinashgeda@ufl.edu, Tharanga Rajapakshe, Asoo Vakharia
We develop a model and solution procedure for preemptive scheduling with
overtime option. The problem is mainly motivated by the aviation MRO industry
where late deliveries of overhaul orders are costly. Our model aims at minimizing
the total cost of overtime and tardiness over a finite number of jobs with different
weights, release dates and due dates. A multi-pass heuristic algorithm is proposed
to solve the scheduling and capacity allocation problem.
3 - Optimal Learning Control of Drone Operation
Jinkun Lee, The Pennsylvania State University, 236 Leonhard,
University Park, PA, United States of America,
jinkunlee@psu.edu, Vittal Prabhu
We consider a monopolist durable goods manufacturer who markets its products
via a single retailer. In the first period, the manufacturer introduces first version
of the product while he may introduce the second and refurbished versions of the
product in the second period. We consider there exists a secondary market where
consumers can resell the old products. We investigate the impact of introduction
of a trade-in and refurbishing program on the product versioning decision of the
manufacturer.
An optimal learning control of individual delivery drone has been considered.
This enables each drone to adapt its path for the minimum travel time according
to its own repetitive experience of the dynamic environment. The estimated
travel times of drones are fed back into the central server, and this server
determines the proper number of drones to fulfill the delivery service level based
on the accumulated real time demands during the previous drone operation time
window.
2 - Trade-Ins Versus Upgrades: A Behavioral Exploration
Mahdi Mahmoudzadeh, Georgia Institute of Technology,
800 W Peachtree ST NW, Atlanta, GA, United States of America,
Mahdi.Mahmoudzadeh@scheller.gatech.edu, Beril Toktay,
Basak Kalkanci
4 - Integrating Taxi Planning and Gate Assignment
Angel Marin, Professor, Polytechnical University of Madrid, ETSI
Aeronautica y Del Espacio, Plaza Cardenal Cisneros, 3, Madrid,
MA, 28040, Spain, angel.marin@upm.es
Understanding customers’ behavior in selling positions or exchanging their
products would help better manage replacement purchases and product return
streams. We study trade-ins and upgrades, which so far have been assumed to be
equivalent. We find that customers perceive trade-ins and upgrades differently;
perceived importance of the quoted price for current product is more salient in
trade-ins than in upgrades. Our results are useful to find dominant replacement
offers and pricing strategies.
In the presentation is studied the Taxi Planning and Gate Assignment Integration.
These problems are considered under a binary multicommodity network flow, in
the context of routing and scheduling models with additional Side Constraints
(SC). The model is a multiobjective approach balancing conflictive objectives:
airport throughput, travel time, delays, operation safety and costs, etc. The
computational tests are realized on test airports, simulating actual ones.
3 - Design and Technology Choice for Recycling: The Value of
Collaboration and Capacity Ownership
Luyi Gui, Assistant Professor, UC Irvine, United States of America,
luyig@exchange.uci.edu, Morvarid Rahmani, Atalay Atasu
Efficient and effective treatment of end-of-life products requires not only product
design improvements but also advancement in recycling technologies. We analyze
how Extended Producer Responsibility (EPR) legislation would affect incentives
for improving product recyclability and processing technology. In particular, we
take into account the mutually reinforcing effect between product and process
improvements and explore the implication of such complementarity in EPR
implementation.
305
TB64
■ TB64
2 - How Little Do Models Tell Us?
Eva Regnier, Associate Professor, Naval Postgraduate School,
699 Dyer Road, Monterey, CA, 93943, United States of America,
eregnier@nps.edu, Erin Baker
64-Room 113A, CC
Panel Discusssion: A Heated Discussion on Decision
Analysis and Systems Engineering
In arenas including weather forecasting and climate policy, simulation modeling is
used to estimate uncertainty attributed to initial conditions. Model uncertainty
(sometimes called structural uncertainty) is much harder to quantify. We outline
a qualitative approach using Bayesian logic to answer the question: how much do
model results tell us?
Sponsored Session
Chair: Ali Abbas, Professor Of Industrial And Systems Engineering And
Public Policy And Director Of Create, University of Southern California,
3710 McClintock Avenue, RTH 314, Los Angeles, CA,
United States of America, aliabbas@price.usc.edu
1 - The Need for a Sound Decision Making System
Moderator:Ali Abbas, Professor Of Industrial And Systems
Engineering And Public Policy And Director Of Create, University
of Southern California, 3710 McClintock Avenue, RTH 314, Los
Angeles, CA, United States of America, aliabbas@price.usc.edu
3 - Agile Modeling Focused on Decision Making
Max Henrion, CEO, Lumina Decision Systems, Inc, 26010
Highland Way, Los Gatos, CA, 95033, United States of America,
henrion@lumina.com
Agile modeling borrows methods from agile software development, an alternative
to the conventional approaches starting from formal requirements. Instead
modelers start building a simple prototype, and refine it progressively, learning
and improving as they go. Decision analysis and sensitivity analysis helps focus
development on areas most decision-relevant.
This talk reflects on some widely used methods of multi-objective decision
making in both public and private enterprises, and demonstrates the issues with
their use and the need for a sound decision making system.
■ TB66
2 - Ethical Decision Analysis
Ronald Howard, Professor, Stanford University, 646 Tennyson
Avenue, Palo Alto, CA, 94301, United States of America,
rhoward@stanford.edu
66-Room 113C, CC
Delay Propagation and Robust Airline Operations
Sponsored Session
Decision analysis is inherently amoral. Like fire or nuclear energy it can be used
for good or ill. The decision analyst and the decision maker have the ethical
responsibility for decisions. The decision maker for the choice of action and the
decision analyst as a conspirator or accomplice in clarifying what is to be done.
The daily news shows the consequences of abdicating ethical responsibility.
Chair: Milind Sohoni, Associate Professor Of Operations Management
And Sr. Associate Dean Of Programs, Indian School of Business,
Gachibowli, Indian School of Business, Gachibowli, Hyderabad, Pl,
500032, India, milind_sohoni@isb.edu
1 - Improving Maintenance Robustness using a Route Adjustment
Tail Assignment Problem
Stephen Maher, Zuse Institute Berlin, Takustr. 7, Berlin, BE,
14195, Germany, maher@zib.de, Guy Desaulniers,
François Soumis
3 - There is No Rational Framework for Systems Engineering
George Hazelrigg, Deputy Division Director, National Science
Foundation, Civil, Mech. & Mfg Innovation, 4201 Wilson
Boulevard, Arlington, VA, 22230, United States of America,
ghazelri@nsf.gov
Decision analysis for systems engineering is an oxymoron. Systems engineering
requires teams of people, for which decision analysis does not apply. Failure to
recognize this can lead to serious problems.
Maintenance planning is critical for airline operations. Daily schedule
perturbations regularly prohibit aircraft from receiving maintenance as required.
A robust approach employing one-day routes has been proposed, however,
perturbations still affect the delivery of maintenance. A tail assignment problem
that modifies routes to satisfy maintenance requirements is presented. This will
demonstrate that route modifications are a necessary augmentation to a robust
maintenance planning solution.
4 - Decision Analysis for Systems Engineering Trade-off Analyses
Greg Parnell, Professor, University of Arkansas, Department of
Industrial Engineering, Fayetteville, AR, 72701,
United States of America, gparnell@uark.edu
Critical systems decisions are made throughout the system life cycle. Decision
analysis offers a sound foundation for developing a composite model of complex
system alternatives, major uncertainties, and stakeholder values to provide
insights to systems decision makers.
2 - Examining the Robustness of Airline Operations under
Weather Disruptions
Donald Richardson, University of Michigan, Ann Arbor, MI,
donalric@umich.edu, Luke Stumpos, George Tam, Amy Cohn,
Chhavi Chaudhry
5 - Decision Analysis - Towards a Theoretical Foundation of Systems
Engineering and Design
Chris Paredis, Program Director, National Science Foundation,
4201 Wilson Blvd, Arlington, VA, United States of America,
cparedis@nsf.gov
We have compiled a database containing twelve years’ worth of flight data from
the Bureau of Transportation Statistics. By connecting this data with hourly
National Oceanic and Atmospheric Administration weather reports, we are able
to analyze how the weather affects the relationship between planned airline
schedules and the actual flight performance. The purpose of this research is to
provide a foundation for better understanding the robustness of airline operations
under weather disruptions.
In a rapidly changing global context, out approach for engineering large-scale,
complex engineered systems must also adapt quickly. A theoretical foundation for
systems engineering and design is needed to help guide this adaptation in a
rigorous, systematic fashion. Decision analysis is an important cornerstone of this
foundation.
3 - Data-driven Models for Robust Aircraft Routing
Lavanya Marla, Assistant Professor, University of Illinois at
Urbana-Champaign, 104 S. Mathews Avenue, 216E, Urbana, IL,
61801, United States of America, lavanyam@illinois.edu,
Vikrant Vaze
■ TB65
65-Room 113B, CC
We address the issue of pro-actively building robust aircraft routings that are less
vulnerable to uncertainty, by focusing on reducing delay propagation. We present
a series of data-driven models drawn from the classes of Robust Optimization and
Chance-Constrained Programming that generate solutions that (i)are faithful to
implicit information in the underlying data, and (ii)are less fragile to disruption.
We conclude with results from a real-world airline network to provide proof-ofconcept.
Modeling in Decision Analysis
Sponsored Session
Chair: Jeffrey Keisler, University of Massachusetts Boston,
100 Morrissey Boulevard, Boston, MA, 02125,
United States of America, Jeff.Keisler@umb.edu
1 - When Decision Analysis Serves to Connect a Network
Jeffrey Keisler, University of Massachusetts Boston,
An organization may wish to construct analytic models combining contributions
from different experts and stakeholders in order to guide decisions. We represent
this as a network of agents with reporting relationships, each with a vocabulary, a
knowledge base, potential observations. Is the network rich enough to ensure the
decider’s success? Recent results from mathematical logic give some answers and
possible implications for decision consulting.
306
TB69
■ TB68
4 - Can Time Buffers Lead to Delays? The Role of
Operational Flexibility
Milind Sohoni, Associate Professor Of Operations Management
And Sr. Associate Dean Of Programs, Indian School of Business,
Gachibowli, Indian School of Business, Gachibowli, Hyderabad,
Pl, 500032, India, milind_sohoni@isb.edu, Sanjiv Erat
68-Room 201B, CC
TSL Invited Cluster Keynote Address
Sponsored Session
In operating systems where the feasible start time of activities is uncertain, the
actual buffers for conducting the activities are distinct from scheduled buffers. We
study how, and why, do these buffers affect operating performance? We propose a
theoretical model and evaluate its empirical content and predictions using airline
industry data. Our main result shows that both buffers impact performance and
their effects are moderated by flexibility. Thus ex-ante plans must consider
flexibility.
Chair: Irina Dolinskaya, Northwestern University, 2145 Sheridan Road,
Evanston, IL, 60208, United States of America,
dolira@northwestern.edu
1 - Stochastic Vehicle Routing: An Overview and Some
Research Directions
Michel Gendreau, Full Professor, École Polytechnique de
Montréal, P.O. Box 6128, Station Centre-ville, Montreal, QC, H3C
3J7, Canada, michel.gendreau@cirrelt.ca
■ TB67
While Vehicle Routing Problems have now been studied extensively for more
than 50 years, those in which some parameters are uncertain at the time where
the routes are made have received significantly less attention, in spite of the fact
that there are many real-life settings where key parameters are not known with
certainty. In this talk, we will examine the main classes of Stochastic Vehicle
Routing Problems: problems with stochastic demands, stochastic customers, and
stochastic service or travel times. We will emphasize the main approaches for
modeling and tackling uncertainty: a priori models, a posteriori approaches,
and chance-constrained models. The end of the talk will devoted to a brief
presentation of some interesting research directions in this area.
67-Room 201A, CC
Advances in Vehicle Routing Problem and its Variants
Sponsored Session
Chair: Ibrahim Capar, The University of Alabama, Box 870226,
Tuscaloosa, AL, United States of America, icapar@cba.ua.edu
1 - The Vehicle Routing Problem with Drones: A Worst-case Analysis
Xingyin Wang, University of Maryland, Mathematics department,
University of Maryland, College Park, MD, 20742, United States
of America, wangxy@umd.edu, Stefan Poikonen, Bruce Golden
■ TB69
We introduce the Vehicle Routing Problem with Drones (VRPD). A fleet of trucks
equipped with drones delivers packages to customers. Drones can be dispatched
from and picked up by the trucks at the depot or the customer locations. The
objective is to minimize the maximum duration of the routes. We compare VRPD
to the min-max Vehicle Routing Problem from a worst-case perspective and show
that the maximum savings from using the drones depends on the number and the
speed of the drones.
69-Room 201C, CC
Joint Session TSL/Public Sector: Health-care,
Education, and Emergency Applications of Logistics
Sponsored Session
2 - Online and Open Vehicle Routing Problem with Split Delivery
Ibrahim Capar, The University of Alabama, Box 870226,
Tuscaloosa, AL, United States of America, icapar@cba.ua.edu,
Burcu Keskin
Chair: Sung Hoon Chung, Binghamton University, P.O. Box 6000,
Binghamton, NY, United States of America, schung@binghamton.edu
1 - Public Transportation Planning for Mass-Scale Evacuations
Rahul Swamy, Graduate Research Assistant, University at Buffalo
(SUNY), 412 Bell Hall, Buffalo, NY, Jee Eun Kang, Rajan Batta
We consider an online, open vehicle routing problem with split deliveries. This
type of problem is usual for companies that use common carriers with TL, LTL, or
container services. We develop an integer programming model and propose a
reduction technique to solve real life problem with commercial software. We
investigate the effect of lead time on cost and outstanding orders and explore
different policies to minimize total cost. We show more than eight percent savings
compared to the literature.
This research provides a public transportation planning strategy in an urban
setting for evacuating population groups to safe locations before a mass-scale
disaster. Under the objective of maximizing the number of evacuees, the proposed
model first identifies pickup locations and then constructs special type of routing
to serve a time-varying demand.
2 - Minimizing the Cost of Routing Blood Collection Vehicles
Okan Orsan Ozener, Ozyegin University, Cekmekoy, Istanbul,
Turkey, orsan.ozener@ozyegin.edu.tr
3 - Distributionally Robust Adaptive Vehicle Routing
Arthur Flajolet, MIT, Operations Research Center,
77 Massachusetts Avenue, Bldg. E40-149, Cambridge, MA,
02139, United States of America, flajolet@mit.edu, Patrick Jaillet,
Sebastien Blandin
We study the routing of blood collection vehicles to minimize the total routing
costs. Donated blood has to be processed within a certain amount of time. We
analyze the routing decisions and propose an integrated framework to minimize
the total cost while collecting a pre-specified number of donations.
We consider an adaptive solution to the vehicle routing problem with stochastic
travel times with the objective of minimizing a risk function of the lateness. To
mitigate the impact of the lack of information on the travel times, we develop a
distributionally robust dynamic programming formulation for risk-averse travelers
and illustrate the practicality of the approach with field data from the Singapore
road network.
3 - A Heuristic for School Bus Routing of Special-education Students
Hernan Caceres, SUNY Buffalo, 342 Bell Hall, Buffalo, NY, United
States of America, hernanan@buffalo.edu, Rajan Batta, Qing He
The problem of routing special-education students differs in many aspects with
that of routing regular students. A bus can be configured to also support
wheelchairs, students may be served differently depending on their disability, and
they need to be picked up and dropped off in their homes. In our study we
modeled a mixed integer program that accounts for these and other
characteristics. We use column generation to find approximated solutions for real
and benchmark instances.
4 - A Metaheuristic for the Electric Vehicle Routing Problem with
Recharging Stations and Time Windows
Site Wang, Graduate Student, Clemson University, 854
Issaqueena Trail, APT908, Central, SC, 29630, United States of
America, sitew@clemson.edu, Eric Huang, Scott Mason
In this study, we consider electric vehicles and recharging stations in the vehicle
routing problem with time windows. We examine two objectives for this problem,
separately and in concert, to provide insights for the location-routing problem
with time windows. Due to the problem’s complexity, we demonstrate the
efficacy of our two-phase metaheuristic that combines variable neighborhood
search and Tabu search for practical-sized problems.
4 - Disaster Relief Routing under Uncertainty: A Robust
Optimization Approach
Sung Hoon Chung, Binghamton University, P.O. Box 6000,
schung@binghamton.edu, Yinglei Li
We explicitly consider uncertainty in travel times when planning vehicle routes
for delivering critical supplies to the affected population in need in the aftermath
of a large disaster. In particular, we propose the robust optimization approach to
minimize the impact of uncertainty and eventually to achieve enhanced resilience
in the aftermath of disasters. We also explore several numerical methods and
algorithms.
307
TB70
■ TB70
quantitative analytic tools. We introduce an automated tool suitable for analyzing
the in-situ TEM videos. It learns and tracks the normalized particle size
distribution and identifies the phase change points delineating the stages in
nanocrystal growth. We furthermore produce a quantitative physical-based
model.
70-Room 202A, CC
Yard and Terminal Simulation
Sponsored Session
2 - Cooperative Unmanned Vehicles for Vision-based Detection and
Real-world Localization of Human Crowds
Sara Minaeian, The University of Arizona, 1127 E James E.
Rogers Way, Room 111, Tucson, AZ, 85716, United States of
America, minaeian@email.arizona.edu, Young-jun Son, Jian Liu
Chair: Roger Baugher, President, TrAnalytics, LLC, 100 Villamoura
Way, Johns Creek, GA, 30097, United States of America,
rwbaugher@aol.com
1 - Exploiting Data to Create Yard and Terminal Replay Capabilities
Roger Baugher, President, TrAnalytics, LLC, 100 Villamoura Way,
Johns Creek, GA, 30097, United States of America,
rwbaugher@aol.com
In crowd control using unmanned vehicles (UVs), the crowd detection and realworld localization are required to perform key functions such as tracking and
motion planning. In this work, a team of UVs cooperates under a DDDAMS
framework to detect the moving crowds by applying computer-vision techniques
and to localize them using a new perspective transformation. A simulation model
is also developed for validation, and the experimental results reveal the
effectiveness of the proposed approach.
Yard automation technology, GPS sensors, time lapse cameras and new low cost
computer processors enable large amounts of yard operation data to be captured
inexpensively. Processes can transform these data, and the yard’s GIS data, into
inputs for simulation, enabling the deployment of yard replay systems. With such
a system, management can analyze operational failures, develop improved
processes, train new employees, examine the impact of proposed capital
improvements and more.
3 - Fault Identifiability Analysis of Beam Structures using Dynamic
Data-driven Approaches
Yuhang Liu, Research Assistant, University of WisconsinMadison, 1513 University Ave, ME3255, Madison, WI, 53706,
United States of America, liu427@wisc.edu, Shiyu Zhou
2 - Simulation Model for a Large Railroad Flat Switching Yard
Clark Cheng, Senior Director Operations Research, Norfolk
Southern Railway, Atlanta, GA, 30309, United States of America,
Clark.Cheng@nscorp.com, Rajesh Kalra, Mabby Amouie,
Edward Lin
In this research, we study the parameterization and localization identifiability of
beam structures based on the dynamic response information. We show that the
stiffness parameters can be locally identifiable in general cases for the collocated
single input and single output system. The unique relationship between the
damage location and the dynamic response are also investigated. The identifiable
sensitivity is studied for practical damage identification.
We will present a discrete-event simulation model for the largest railroad flat
switching yard in the Western Hemisphere. The model is being used to evaluate
yard capacity and improve yard operations and customer service.
■ TB73
3 - Conflict Avoidance in Yards and Terminals
Brigitte Jaumard, Professor And Concordia Research Chair On
The Optimization Of Communication Networks, Concordia
University, Computer Science and Software Eng., 1455 de
Maisonneuve Blvd. West, Montreal, QC, H3G 1M8, Canada,
bjaumard@cse.concordia.ca, Roger Baugher, Thai Hoa Le,
Bertrand Simon
73-Room 203B, CC
Joint Session QSR/Energy: Data Analytics in
Energy Systems
Sponsored Session
Activities of a rail yard focus on freight delivery and vehicle maintenance, while
train movements are generally line-of-sight ones. Many of the yard activities
share one or two connecting tracks for through traffic. While these tracks need to
remain clear for through traffic, stopping yard activities on them to let a passenger
train through may result in disruption to freight operations, and in conflicts. We
will propose different mechanisms and tools in order to avoid conflicts.
Chair: Eunshin Byon, Assistant Professor, University of Michigan,
ebyon@umich.edu
Co-Chair: Arash Pourhabib, Assistant Professor, Oklahoma State
University, 322 Engineering North, Stillwater, OK, 74078,
United States of America, arash.pourhabib@okstate.edu
1 - Multi-Component Replacement in a Markov
Modulated Environment
David Abdul-Malak, dta10@pitt.edu, Jeffrey Kharoufeh
4 - Applying Dynamic Simulation to Validate and Improve New
Transloading Terminal Operations
Martin Franklin, Partner, MOSIMTEC LLC, 297 Herndon
Parkway, Suite 301, Herndon, VA, 20170,
United States of America, martin@mosimtec.com
In this talk we will present a model for jointly replacing multiple components that
degrade in a shared, exogenous, Markov modulated environment. Continuous
state variables and a high dimensional state space cause the problem to be
computationally intractable. To overcome this complication, an approximate
dynamic programming (ADP) approach is employed and illustrated through
multiple numerical examples.
A chemical manufacturing and handling company is expanding and reconfiguring facilities to create a new interface point between rail transport and
pipeline transport. The client recognized the need to apply modeling and
simulation technology to represent the system in a dynamic environment, therein
incorporating inherent variability, to validate the design and make informed
decisions. Simulation analysis of the rail network and operators and related
integration will also be reviewed.
2 - Importance Sampling with a Novel Information Criterion for
Efficient Reliability Evaluation
Youngjun Choe, PhD Candidate, University of Michigan, 1205
Beal Avenue, Ann Arbor, MI, 48109, United States of America,
yjchoe@umich.edu, Eunshin Byon
■ TB72
72-Room 203A, CC
Importance sampling can significantly accelerate the rare event probability
estimation. However, the theoretically optimal sampling requires some
approximation in practice, such as the cross-entropy method. We extend the
cross-entropy method by incorporating the expectation-maximization (EM)
algorithm and deriving a model selection criterion analogous to Akaike
information criterion. We apply the proposed method to the reliability evaluation
of the wind turbine.
Applications II
Sponsored Session
3 - Monitoring Performance of Wind Turbines Based on Power
Curve Estimation
Hoon Hwangbo, PhD Student, Texas A&M University, College
Station, TX, United States of America, hhwangbo@tamu.edu,
Andrew Johnson, Yu Ding
MS 3131, College Station, YX, United States of America,
1 - Multi-stage Nanocrystal Growth Identifying and Modeling via
in-situ TEM Video
Yanjun Qian, PhD Candidate, TAMU, 1501 Harvey Rd, Apt. 806,
qianyanjun09@gmail.com, Yu Ding, Jianhua Huang
Quantifying performance of a wind turbine is crucial for decision makings such as
turbine upgrade or replacement. Yet, there is a lack of systematic ways to quantify
a turbine’s performance, while considering the diverse sources of variation in the
energy generation. In this study, we estimate power curves and quantify
performance of a wind turbine while controlling for some significant factors of
variation. Using the measures we derive, we monitor performance change of a
wind turbine over time.
While in-situ transmission electron microscopy technique has caught a lot of
recent attention, one of the bottlenecks appears to be the lack of automated and
308
TB75
■ TB75
4 - Wake Effect Characterization in Wind Power Systems
Mingdi You, PhD Candidate, University of Michigan, 1205 Beal
Avenue, IOE 1773, Ann Arbor, MI, 48109, United States of
America, mingdyou@umich.edu, Eunshin Byon,
Jionghua (judy) Jin
75-Room 204B, CC
IBM Research Best Student Paper Award II
Sponsored Session
The rapid growth of wind power underscores the need to understand the
dynamic characteristics of wind turbine operations. Wind turbines in a wind farm
exhibit heterogeneous power generations due to the wake effect. This study
provides a computational framework for characterizing the wake effects via a
data-driven approach by extending the Gaussian Markov Random field
framework. The computational results show that this approach improves the
prediction capability over other methods.
huangmh@ntu.edu.tw
huangmh@ntu.edu.tw
■ TB74
74-Room 204A, CC
System and Process Informatics in Additive
Manufacturing (II)
2 - Online Network Revenue Management using Thompson
Sampling
He Wang, MIT, Cambridge, MA, United States of America,
wanghe@mit.edu, Kris Johnson Ferreira, David Simchi-Levi
Sponsored Session
Mobile apps have great potential to provide promising services to improve
consumers’ engagement and behaviors. Focusing on healthy eating, this study
shows that an image-based professional support greatly improves consumer
engagement and eating behaviors, while social media and a heuristic approach of
self-management might have negative effects in some occasions. Mobile apps
have great potential to provide promising services to improve consumers’
engagement and behaviors. Focusing on healthy eating, this study shows that an
image-based professional support greatly improves consumer engagement and
eating behaviors, while social media and a heuristic approach of self-management
might have negative effects in some occasions.
Chair: Linkan Bian, Assistant Professor, Mississippi State University,
260 McCain Building, Mississippi State, Starkville, MS, 39762,
United States of America, bian@ise.msstate.edu
1 - Accelerated Bi-objective Process Optimization for Laser-based
Additive Manufacturing (LBAM)
Amir M. Aboutaleb, Mississippi State University, 260 McCain
Building, Mississippi State, MS, 39762, United States of America,
aa1869@msstate.edu, Alaa Elwany, Scott M. Thompson,
Linkan Bian, Nima Shamsaei, Mohammad Marufuzzaman
3 - How Environmental Certification Can Affect Performance in the
Service Industry: Evidence from the Adoption of LEED Standards
in the U.S. Hotel Industry
Matthew Walsman,Cornell University, Ithaca, NY United States
of America, mcw237@cornell.edu , Suresh Muthulingam,
Rohit Verma
Material properties of fabricated parts via LBAM have demonstrated to either be
correlated, interdependent or inconsistent with process parameters. In many cases
the goal is optimize the LBAM process considering several material properties of
interest as a multi-objective problem. We propose a novel methodology for
leveraging current experimental data to guide and accelerate the bi-objective
Design-of-Experiment process for Pareto Front approximation by the minimum
number of experiments.
This study uses a mixed method approach (difference-in-differences and multilevel modeling) to measure the impact of environmental certification (i.e. LEED
certification) on financial performance in the US hospitality industry. We find
that certification does contribute to higher revenue for the certifying hotel,
relative to its competitors.
2 - Spatial Gaussian Process Models for Porosity Prediction in
Selective Laser Melting
Alaa Elwany, Texas A&M University, 3131 TAMU, College
Station, TX, United States of America, elwany@tamu.edu,
Gustavo Tapia, Huiyan Sang
4 - Optimal Coinsurance Rates for a Heterogeneous Population
under Inequality and Resource Constraints
Greggory J. Schell, Center for Naval Analyses, 3003 Washington
Blvd, Arlington, VA, United States of America, schellg@cna.org,
Rodney A. Hayward, Mariel Lavieri, Jeremy B. Sussman
We develop a Gaussian process-based predictive model for predicting the porosity
in metallic parts produced using Selective Laser Melting (SLM – a laser-based AM
process). A case study is conducted to validate this predictive framework through
predicting the porosity of 17-4 PH stainless steel manufacturing on a commercial
SLM system.
We derive prescription coinsurance rates which maximize the health of a heterogeneous patient population. We analyze the problem as a bilevel optimization
model where the lower level is a Markov decision process and the upper level is
a resource allocation problem with constraints on expenditures and coinsurance
inequality.
3 - Automatic Feature Priority Assignment for Automated
Production Processes
Ola Harryson, Professor, North Carolina State University,
400 Daniels Hall, 111 Lampe Dr, Raleigh, NC, 27606,
United States of America, oaharrys@ncsu.edu, Richard Wysk,
Sidharth Chaturvedi, Harshad Srinivasan
5 - Managing Rentals with Usage-Based Loss
Vincent Slaugh, Penn State University, Univeristy Park, PA,
United States of America, vslaug@cmu.edu, Bahar Biller,
Sridhar Tayur
This work describes a system for the prioritization of features at the near-net
production stage in order to minimize the effort required for any subsequent
finish machining. Heuristics are used to assign weights to features based on value
and produceability. A graph of feature relationships is is used to modify the
assigned weights based on design and tolerancing principles. An implementation
of this system for use with the AIMS hybrid process is described and
demonstrated with sample parts.
We study the operation of a discrete-time stochastic rental system over a single
selling season in which rental units may be purchased or damaged by customers.
We provide structural results related to the expected profit function and the
optimal policy for allocating rental units to meet customer demand. In an
industrial use case motivated by a high-fashion dress rental business, we show
significant value to accounting for inventory loss and using the optimal inventory
recirculation rule.
4 - Additive Manufacturing of Biomedical Implants:
Feasibility Assessment via Supply-chain Cost Analysis
Adindu Emelogu, Mississippi State University, 260 McCain
Building, Mississippi State, MS, 39762, United States of America,
emeloguadindu@yahoo.com, Linkan Bian,
Mohammad Marufuzzaman
We investigate the economic feasibility of fabricating biomedical implants close to
hospitals by additive manufacturing (AM) instead of traditional manufacturers
(TM) located far from point-of-use. We develop a stochastic mixed-integer
programming model which helps to decide the location of AM centers and
volume of product flows that minimize supply chain cost. A case study of
hospitals in Mississippi, USA recommends AM only when the production cost of
AM to TM ratio (ATR) reduces to 3 or less.
309
TB76
■ TB76
2 - Flexibility Analysis on a Supply Chain Contract using a Parametric
Linear Programming Model
Eric Longomo, PhD student, University of Portsmouth, Lion Gate
Building, Lion Terrace, Hampshire, Portsmouth, PO1 3HF,
United Kingdom, eric.longomo@port.ac.uk, Xiang Song,
Djamila Ouelhadj, Chengbin Chu
76-Room 204C, CC
Advances in Simulation-based Optimization II
Sponsor: Simulation
Sponsored Session
This study considers a multi-period Quantity Flexibility contract between a car
manufacturer (buyer) and an external parts supplying company. The buyer -in
concert with the supplier- aims to develop a policy –at strategic level, that
determines the optimal nominal order quantity and variation rate underpinning
the contract. The feasibility and convexity of the proposed LP model are
examined. Simulations are carried out to evaluate the theoretical results.
Chair: Enlu Zhou, Assistant Professor, Georgia Institute of technology,
755 Ferst Drive, NW, Atlanta, United States of America,
enlu.zhou@isye.gatech.edu
1 - A Set Approach to Simulation Optimization with Probabilistic
Branch and Bound
Hao Huang, PhD Candidate, University of Washington, Industrial
and Systems Engineering, Seattle, WA, 98195-2650,
United States of America, haoh7493@uw.edu, Zelda Zabinsky
3 - Assigning Non-Fixed Parts of a Delivery Area to Fixed Tours
Serviced by Electric Vehicles
Sarah Ubber, RWTH Aachen University, Kackertstrafle 7,
Aachen, Germany, ubber@dpor.rwth-aachen.de
Probabilistic Branch and Bound (PBnB) is a partition-based random search
simulation optimization algorithm for stochastic problems. PBnB determines a set
of solutions through an estimated bound on the performance. For single objective
problem, PBnB approximates a desirable level set with quantile estimation. In a
multiple objective circumstance, PBnB considers a bound of the closeness to the
efficient frontier and approximates the Pareto optimal set of solutions.
We consider last mile distribution where a delivery area is operated by different
tours. Parts of this area are serviced by fixed tours in a fixed sequence every day.
Other parts are not assigned to fixed tours. To respond e.g. to variable battery
ranges or to fluctuations in demand, it is useful to reassign daily the non-fixed
parts to the tours, whereby the assignment must not significantly alter the usual
delivery sequence. We have developed a model and a heuristic for solving this
problem.
2 - A Model-based Approach to Multi-objective Optimization
Joshua Hale, Graduate Student, Georgia Institute of Technology,
755 Ferst Drive, NW, Atlanta, GA, Atlanta, GA, 30332,
United States of America, jhale32@gatech.edu, Enlu Zhou
4 - Asset Allocation in the Industrial Gas Bulk Supply Chain
Leily Farrokhvar, Virginia Tech, 250 Durham Hall (0118),
Blacksburg, VA, 24061, United States of America, leily@vt.edu,
Kimberly Ellis
We develop a model-based algorithm for the optimization of multiple objective
functions that can only be assessed through black-box evaluation. The algorithm
iteratively generates candidate solutions from a mixture distribution over the
solution space and updates the mixture distribution based on the sampled
solutions’ domination count such that the future search is biased towards the set
of Pareto optimal solutions. We demonstrate the performance of the proposed
algorithm on benchmark problems.
We study an asset allocation problem in a vendor managed inventory system of
an industrial gas distribution network where customer demands vary over time.
The objective is to determine the preferred size of bulk tanks to assign to customer
sites to minimize recurring gas distribution costs and initial tank installation costs
while accommodating customers’ time varying demand. The problem is modeled
as a mixed-integer program and then solved using a periodically restricting
heuristic approach.
3 - Simulation Optimization: Review and Exploration
Chun-hung Chen, George Mason University, 4400 University
Drive, MS 4A6, SEOR Dept, GMU, Fairfax, VA, 22030, United
States of America, cchen9@gmu.edu, Edward Huang, Jie Xu,
Loo Hay Lee
■ TB78
78-Room 301, CC
Recent advances in simulation optimization research and explosive growth in
computing power have made it possible to optimize complex stochastic systems
that are otherwise intractable. We will review some recent developments. We will
also discuss how simulation optimization can benefit from cloud computing and
high-performance computing, its integration with big data analytics, and the
value of simulation optimization to help address challenges in engineering design
of complex systems.
Planning and Scheduling in Energy Applications
Contributed Session
Chair: Yanyi He, Senior Scientist, IBM, 1001 E Hillsdale Blvd, Foster
City, Ca, 94404, United States of America, heyanyidaodao@gmail.com
1 - Stochastic and Robust Optimization of the Scheduling and
Market Involvement for an Energy Producer
Ricardo Lima, KAUST, Thuwal, Thuwal, Saudi Arabia,
ricardo.lima@kaust.edu.sa, Sabique Langodan, Ibrahim Hoteit,
Antonio Conejo, Omar Knio
4 - MO-MO2TOS for Multi Objective Multi Fidelity
Simulation Optimization
Loo Hay Lee, National University of Singapore,
Department of Industrial & Systems, Engineering, Singapore,
iseleelh@nus.edu.sg, Giulia Pedrielli, Chun-hung Chen,
Ek Peng Chew, Haobin Li
We will present three optimization methods based on stochastic programming,
robust optimization, and a hybrid method for the scheduling and market
involvement for an electricity producer. This producer operates a system with
thermal, hydro, and wind sources. The wind power and the electricity prices are
uncertain. The methods are implemented using parallel optimization runs. The
computational performance, scheduling results, and the impact of risk
management are presented and discussed.
In simulation–optimization, low fidelity models can be particularly useful.
However, we need to account for their inaccuracy while searching for the
optimum. In 2015, Xu et al. proposed MO2TOS, which exploits multiple fidelities
to improve the simulation optimization procedure. We extend the approach
proposing MO–MO2TOS for the multi-objective case, using the concepts of
non–dominated sorting and crowding distance. Several interesting insights
specific to the multi-objective case are drawn.
2 - A Two-Echelon Wind Farm Layout Planning Model
Huan Long, City University of Hong Kong, Room 601,
Nam Shan Estat,, Hong Kong, China, hlong5-c@my.cityu.edu.hk,
Zijun Zhang
■ TB77
In this paper,a two-echelon layout planning model is proposed to determine the
optimal wind farm layout to maximize its expected power output.In the first
echelon,a grid composed of identical cells is utilized to model the wind farm while
the cell center is the potential slot.In the second echelon, the model for
determining the optimal coordinate in a grid cell is formulated.The comparative
analysis between the two-echelon planning model and the traditional
grid/coordinate models is conducted.
77-Room 300, CC
Logistics I
Contributed Session
Chair: Leily Farrokhvar, Virginia Tech, 250 Durham Hall (0118),
Blacksburg, VA, 24061, United States of America, leily@vt.edu
1 - Analysis of a New Dual-Command Operation in Puzzle-Based
Storage Systems with Block Movement
Hu Yu, PhD Student, University of Science and Technology of
China, Number 96, JinZhai Road, HeFei, China,
yuhu0421@mail.ustc.edu.cn, Yugang Yu
3 - Demand Side Participation for a Major Consumer in a
Co-optimized Electricity and Reserve Markets
Mahbubeh Habibian, Miss, University of Auckland,
6A-Short St, Auckland Central, Auckland, 1010, New Zealand,
mhab735@aucklanduni.ac.nz, Golbon Zakeri,
Anthony Downward
Dual-command operation jointly performing storage and retrieval requests has
been widely discussed in classical warehouse systems, but has been rarely studied
in puzzle-based storage systems with block movement. We analytically derive the
travel time of completing dual requests that randomly locate in the system.
Comparison results with traditional dual-command operation in different
scenarios show that significant reduction in the expected travel time is obtained
in puzzle-based systems.
The paper probes demand side participation for a large consumer through
demand response and offering in interruptible load reserve. Our model is a bilevel optimization problem that embeds the dispatch model, where electricity and
reserve are co-optimized, as the lower level and the profit maximization problem
for the consumer (over 2 sets of supply functions) as the upper level. The
objective function is transformed into piecewise linear form via utilizing a new
interpretation of offer stacks.
310
4 - Analysis of Best Practices for Energy Efficient Buildings through
Building Energy Modeling in Design
Chung-suk Cho, Assistant Professor, Khalifa University, Al Saada
St. and Muroor Rd., Abu Dhabi, 127788, United Arab Emirates,
chung.cho@kustar.ac.ae, Young-ji Byon
POSTER SESSION
model.
2 - Inventory Control with Unknown Demand and
Nonperishable Product
Tingting Zhou, Rutgers University, 1 Washington Park, Newark,
NJ, 07102, United States of America, tingzhou@rutgers.edu,
Michael Katehakis, Jian Yang
Building energy performance modeling evaluates energy efficient design options.
There are significant amount of misused opportunities of energy efficiency-related
decisions that could be incorporated into the building design through quality
building energy performance modeling. The best practice analysis will help
optimize the building design and will allow the design team to prioritize
investment in the strategies that will have the greatest effect on the building’s
energy use.
We study an inventory control problem with unknown discrete demand
distribution, focusing on the analysis of an adaptive algorithm based on empirical
distributions and the newsvendor formula. When items are nonperishable, the
algorithm can achieve a near square-root-of-T bound on its regret over the ideal
case where demand distribution were known.
3 - Optimizing Information System Security Investments with Risk:
Insights for Resource Allocation
Yueran Zhuo, PhD Candidate, University of Massachusetts
Amherst, Isenberg School of Management, Amherst, MA, 01003,
United States of America, yzhuo@som.umass.edu, Senay Solak
5 - Two-stage Optimal Demand Response with Battery Energy
Storage Systems
Yanyi He, Senior Scientist, IBM, 1001 E Hillsdale Blvd,
Foster City, CA, 94404, United States of America,
heyanyidaodao@gmail.com, Zhaoyu Wang
Information security has become an integral component of a firm’s business
success, and thus investing on information security countermeasures is an
important decision problem for many businesses. We use a portfolio approach to
study the optimal investment decisions of a firm, where the uncertainty of
information security environment is captured through a stochastic programming
framework. Results cast managerial insights for information security investment
planning by a firm.
Proposes a two-stage co-optimization framework for the planning and energy
management of a customer with battery energy storage systems (BESSs) and
demand response (DR) programs. The first stage is to assist a customer to select
multiple DR programs to participate and install batteries to coordinate with the
demand side management. The second stage is to perform energy management
according the planning decisions, including dispatches of batteries, loads, and
DGs.
4 - Adaptive Decision-Making of Breast Cancer Mammography
Screening: A Heuristic-Based Regression Model
Fan Wang, University of Arkansas,, Fayettevlle, AR,
United States of America, fxw005@uark.edu, Shengfan Zhang
■ TB79
79-Room 302, CC
The American Cancer Society currently recommends all U.S. women undergo
routine mammography screenings beginning at age 40. However, due to the
potential harms associated with screening mammography, such as overdiagnosis
and unnecessary work-ups, the best strategy to design an appropriate breast
cancer mammography screening schedule remains controversial. This study
presents a mammography screening decision model that aims to identify an
adaptive screening strategy while considering disadvantages of mammography.
We present a two-stage decision framework: (1) age- specific breast cancer risk
estimation, and (2) annual mammography screening decision-making based on
the estimated risk. The results suggest that the optimal combinations of
independent variables used in risk estimation are not the same across age groups.
Our optimal decisions outperform the existing mammography screening
guidelines in terms of the average loss of life expectancy. While most earlier
studies improved the breast cancer screening decisions by offering lifetime
screening schedules, our proposed model provides an adaptive screening decision
aid by age. Since whether a woman should receive a mammogram is determined
based on her breast cancer risk at her current age, our “on-line” screening policy
is adaptive to a woman’s latest health status, which causes less bias in reflecting
the individual risk of every woman.
Invited Session
1 - Statistics.com - A Survey of Data Analytics Methods
Peter Bruce, Founder and President, The Institute for Statistics
Education at Statistics.com
This workshop will survey the field of data analytics, reviewing both traditional
statistical methods and machine learning methods, including predictive modeling,
unsupervised learning, text mining, statistical inference, time series forecasting,
recommender systems, network analytics, and more. It will be a broad brush
treatment aimed at newcomers, as well as those with knowledge in one area who
wish to understand where other analytic methods fit into the picture.
2 - River Logic - Code-free modeling for large-scale LP and MIP
problems using Enterprise Optimizer
Eric Kelso, VP Product Management, River Logic
Enterprise Optimizer is a code-free, visual LP and MIP optimization modeling
platform. Using EO’s intuitive drag-and-drop interface, learn how to rapidly
create integrated process and financial models. Also learn about EO’s wizarddriven data integration, query designer, user-defined schema, dashboard builder,
VBA integration, APIs and job automation component. Outputs demonstrated
include detailed unit costs and audit-quality P&L, Balance Sheet and Cash Flow
statements. The entire session will be spent discussing major features and
showing real-world applications.
5 - Optimization of Netting Scheme in Large-scale Payment Network
Shuzhen Chen, University of Science & Technology of China,
No. 98, Jinzhai Road, Hefei, China, csz@mail.ustc.edu.cn
As netting becomes combined with real-time settlement, an efficient netting
method is required to deal with the large-scale payment network. Network
optimization may not be optimal due to repeated searching of shortest path. A
new method is proposed to optimize the netting process by assembling payments
in two specific routes. It can minimize the amount of total payments for the
whole network and ensure unchanged net payment for each bank. Moreover, it
has polynomial time-complexity.
Tuesday, 12:30pm - 2:30pm
6 - Wasserstein Metric and the Distributionally Robust TSP
Mehdi Behroozi, University of Minnesota, Minneapolis, MN,
United States of America, behro040@umn.edu,
John Gunnar Carlsson
Exhibit Hall A
Tuesday Poster Session
Contributed Session
Recent research on the robust and stochastic travelling salesman problem and the
vehicle routing problem has seen many di?erent approaches for describing the
region of uncertainty, such as taking convex combinations of observed demand
vectors or imposing constraints on the moments of the spatial demand
distribution. One approach that has been used outside the transportation sector is
the use of statistical metrics that describe a distance function between two
probability distributions. In this paper, we consider a distributionally robust
version of the Euclidean travelling salesman problem in which we compute the
worst-case spatial distribution of demand against all distributions whose earth
mover’s distance to an observed demand distribution is bounded from above. This
constraint allows us to circumvent common overestimation that arises when
other procedures are used, such as fixing the center of mass and the covariance
matrix of the distribution.
Chair: Min Wang, Drexel University, 3141 Chestnut Street,
Philadelphia, PA, United States of America, mw638@drexel.edu
Co-Chair: Allen Holder, Rose-Hulman Mathematics, Terre Haute, IN,
United States of America, holder@rose-hulman.edu
Co-Chair: Wenjing Shen, Drexel University, Philadelphia, PA,
United States of America, ws84@drexel.edu
1 - Surgery Scheduling with Recovery Resources
Maya Bam, University of Michigan, Industrial and Operations
Engineering, 1205 Beal Ave., Ann Arbor, MI, 48109,
United States of America, mbam@umich.edu, Mark Van Oyen,
Mark Cowen, Brian Denton
Surgery scheduling is complicated by the post-anesthesia care unit, the typical
recovery resource. Based on collaboration with a hospital, we present a novel, fast
2-phase heuristic that considers both surgery and recovery resources. We show
that each phase of the heuristic has a tight provable worst-case performance
bound. Moreover, the heuristic performs well compared to optimization based
methods when evaluated under uncertainty using a discrete event simulation
311
POSTER SESSION
7 - Intersection of a Tree Network for the Single Refueling Station
Location Problem
Sang Jin Kweon, PhD Student, The Pennsylvania State
University, 310 Leonhard Building, State College, PA, 16802,
United States of America, svk5333@psu.edu
Calibration can be interpreted as a curve to surface matching problem. We
propose a graph-theoretic non-isometric matching approach to solve this problem
using the graph shortest path algorithm in one-dimensional spaces. For higher
dimensional spaces, we introduce the generalized shortest path concept to solve
the matching problem.
An intersection is the vertex whose degree is greater than two in the network. In
this talk, we consider intersections and develops the methodology that determines
the continuous interval of the potential locations for a single alternative-fuel
refueling station on a tree network, with an objective of maximizing the amount
of traffic flows in round trips per time unit captured by the station.
14 - Location and Coverage Models for Preventing Attacks to
Interurban Transportation Networks
Ramón Auad, Associate Professor, Universidad Católica del Norte,
Of. 318, Bldg. Y1, 0610 Angamos Avenue, Antofagasta, 1240000,
Chile, rauad@ucn.cl, Rajan Batta
8 - Intelligent Tutoring Systems: Future Paradigm of
Educational Environments
Alireza Farasat, University at Buffalo (SUNY), 4433 Chestnut
Ridge Rd Apt. 7, Amherst, NY, 14228, United States of America,
afarasat@buffalo.edu, Alexander Nikolaev
We develop a binary integer programming model to solve this problem, whose
objective is to maximize the expected vehicle coverage across the network over a
time horizon, using decomposition heuristics. To introduce a measure of equity,
we propose two sets of time constraints, considering total vehicle coverage,
inequity and network coverage. We explore scalability of the model for
excessively large instances. All of this features are applied to a case study in
Northern Israel.
Educational systems have witnessed a substantial transition from traditional
educational methods mainly using text books, lectures, etc. to newly developed
systems which are artificial intelligent-based systems and personally tailored to
the learners. We have developed a web-based tool, Crowdlearning which
concentrates on creating an intelligent system that learns to interact with students
and motivates them to more actively participate in the learning process by
proposing their own problems.
15 - An Information-based Framework for Incorporating Travel Time
Uncertainty in Transportation Modeling
Jiangbo Yu, University of California, Irvine, 4101 Palo Verde Rd,
Irvine, CA, 92617, United States of America, jiangby@uci.edu,
Jay Jayakrishnan
9 - Optimized Scheduling of Sequential Resource Allocation Systems
Ran Li, PhD Student, Georgia Institute of Technology, 755 Ferst
Drive NW, Atlanta, GA, 30332, United States of America,
rli63@gatech.edu, Spyros Reveliotis
This paper proposes a modeling framework aimed at systematically incorporating
perceived uncertainty into decision making. The model uses theoretically sound
concepts from information theory, communication, and cognitive science.
Potential applications and implications are identified and demonstrated with
examples.
We consider the scheduling problem of allocating finite reusable resources to
concurrent sequential processes. This problem also involves the logical issue of
deadlock avoidance. Our approach is based on the formal model of the
generalized stochastic Petri-net. Special emphasis is placed on the representational
and computational complexity of the proposed methods, which are controlled
through (i) a pertinent (re-)definition of the target policy spaces, and (ii)
simulation optimization.
16 - Database of Identified Poly and Mono ADP-ribosylated Proteins
Charul Agrawal, Undergraduate Student, Indian Institute of
Technology (IIT) Delhi, Room No ED-16, Himadri Hostel, Hauz
Khas, New Delhi, 110016, India, agrawalcharul09@gmail.com
Poly(ADP-ribose) polymerase (PARP) is a family of enzymes with 17 known
members regulating post translational modification of proteins by attaching a
single ADP ribose unit (MARylation) or a chain of ADP ribose (PARylation).In this
study we have attempted to identify all proteins known to be modified by PARPs
and the methods as well as drugs used in such studies. Our study aims to create
the first ever tool for characterizing these modifications.
10 - Operation Research for Data Mining: An Application to
Medical Diagnosis
Shahab Derhami, Auburn University, 3301 Shelby Center,
Auburn, GA, 36849, United States of America,
sderhami@auburn.edu
17 - Configuring Ecommerce Driven Supply Chains in the
FMCG Sector
Stanley Lim, PhD Candidate, Cambridge University, Department
of Engineering, 17 Charles Babbage Road, Cambridge,
United Kingdom, wtsfl2@cam.ac.uk
Fuzzy rule based classification systems (FRBCSs) have been successfully employed
as a data mining technique where the goal is to discover the hidden knowledge in
a data set and develop an accurate classification model. Despite various heuristic
approaches that have been proposed to learn fuzzy rules for these systems, no
exact optimization approach has been developed for this problem. We propose
integer programming models to learn fuzzy rules for a FRBCS used for medical
diagnosis purpose.
Omnichannel has become the engine of growth in retailing. However, it remains
unclear as to how distribution networks should be configured. This research will
shed light through a framework development, and by drawing theories from
supply chain configuration, resource based view, and transaction cost economics.
Case study approach is adopted to identify the critical factors driving operational
choices and seeks to elaborate the relationships between configuration, capability
and performance.
11 - Forecasting Surges in the Hospital Emergency Department (ED)
Alexander Gutfraind, Chief Healthcare Data Scientist, Uptake
Technologies, 600 W. Chicago Avenue, Chicago, IL, 60654, United
States of America, sasha.gutfraind@uptake.com, Nelson Bowers,
Jim Herzog, Madeline Jannotta, Ilan Kreimont,
Adam Mcelhinney
18 - Benchmarking Construction and Improvement Heuristics for
Classification using Markov Blankets
Daniel Gartner, Carnegie Mellon University, 5000 Forbes Avenue,
dgartner@andrew.cmu.edu, Rema Padman
A major hospital system in the Chicago metro area experiences large unexpected
surges in its Emergency Department (ED).
Using five years of ED admissions we predict ED surges and improve scheduling
of staff.
Data indicates the time of arrival, rooming and discharge and acuity. Total arrivals
per day cannot be predicted accurately with epidemiological climatological,
calendar variables but the state of the ED could be predicted 1-4 hours in advance
with high accuracy using VAR methods.
This study examines construction heuristics in connection with a tabu searchbased improvement heuristic for classification in high dimensional data sets.
Using the UCI machine learning data repository containing benchmark instances
in e.g. health care, we evaluate computation times and information about the
evolution of the Markov blanket graphical models in each phase of the heuristics.
We compare the performance of the approaches using evaluation measures such
as classification accuracy.
12 - A New Measure for Testing Independence
Qingcong Yuan, Graduate Student, University of Kentucky, 300
Alumni Drive Apt. 166, Lexington, KY, 40503, United States of
America, qingcong.yuan@uky.edu, Xiangrong Yin
19 - A Sim-heuristic Algorithm for Robust Vehicle Routing Problems
with Stochastic Demand
Abdulwahab Almutairi, Technology, 9 Horizon Building,
Portsmouth, PO4 8EW, United Kingdom,
abdulwahab.m.almutairi@gmail.com
We introduce a new measure for testing independence between two random
vectors. Our measure differs from that of distance covariance, by using expected
conditional difference of characteristic functions. We propose one empirical
version by slicing on one of the random vectors. This empirical measure is based
on certain Euclidean distance. Its properties, asymptotics and applications in
testing independence are discussed. Implementation and Monte Carlo results are
also presented.
We consider the VRPSD in which customers’ demands are stochastic. We propose
to model and solve the VRPSD by developing a robust optimisation model with a
sim-heuristic solution method to minimise the cost while satisfying all demands.
The method combines MCS with CWS in order to efficiently solve the VRPSD
combinatorial optimisation problem. The results is generating very good quality
solutions compared to those in the literature.
13 - Graph Based Non-isometric Curve to Surface Matching for
Local Calibration
Babak Farmanesh, PhD Student, Oklahoma State University,
322 Engineering North, Stillwater, OK, 74078-5016, United
States of America, babak.farmanesh@okstate.edu,
Balabhaskar Balasundaram, Arash Pourhabib
Calibration refers to the process of adjusting parameters of a computer simulation
so that the simulation responses match the corresponding physical responses.
312
20 - Rocket Stage Optimization in Kerbal Space Program
Nathan Arrowsmith, Rochester Institute of Technology,
2800 Butternut Lane, Canandaigua, NY, 14424,
United States of America, nea4305@rit.edu
POSTER SESSION
27 - Simulation Analysis of Chaotic Storage Policies in Amazon Class
Fulfillment Centers
Sanchoy Das, New Jersey Institute of Technology, University
Heights, Newark, NJ, 07102, United States of America,
das@njit.edu, Sevilay Onal
Kerbal Space Program is a space exploration simulation game. Players design,
launch, and fly multi-stage rockets using a variety parts. The performance of these
vehicles is governed by a realistic physics engine. A model was developed which
minimizes the total mass of each rocket stage by choosing motor and fuel tank
combinations which accurately satisfy the Tsiolkovsky Rocket Equation. By
iteratively solving this problem, the lowest mass or least expensive multi-stage
rocket can be determined.
We evaluate storage policies in Amazon Class Fulfillment (ACF) Centers that
primarily serve internet retail. In classical warehouses a SKU is stored in few fixed
locations, no comingling, in bulk volumes and long interval supply. In a chaotic
policy each SKU is stored in any location, comingled, closer to retail volumes and
frequent supply. In an ACF fulfillment time is the primary objective. We use a
simulator model to analyze and present the relative performance for given levels
of workforce.
21 - Investigation of the Effect of Location, Built Environment and
Urban Forms on Customer Satisfaction
Homa Atefyekta, Sharif University of Technology, No.14, 5th St.,
South Piruzan st, Tehran, 1466643479, Iran,
homa.atefyekta@gmail.com, Hamed Ahangari, Hoda Atef Yekta
28 - Spatial-temporal Coverage Evaluation Methodology for
Multi-satellite Embedded Sensors
Monica Maria De Marchi, Dra, Institute for Advanced Studies, Cel
Av Jose Alberto A do Amarante,1, Sao Jose dos Campos, SP,
12228001, Brazil, monica@ieav.cta.br, Osvaldo Catsumi Imamura,
Diogo Maciel Almeida, Maria Jose Pinto
In this study we examine the effect of location factors, urban forms,
transportation accessibilities, and built environment on the customer satisfaction
and business success in restaurant market. We investigated these relationships in
two different geographical areas: the US and Iran by using Yelp and Fidilio data
respectively. The results of this study could be handful for urban policy makers to
improve the urban livability and business entrepreneurs to enhance the odd of
their success.
The intent of this research is to propose an optimized coverage model for satellite
systems and support the decision-making process related to choosing the best
satellites in a scenario of interest. The appropriate satellites are those whose
sensors are able to visualize and identify targets. The decision model proposed
trades off between temporal resolution and the coverage area extension, but also
considers the cost to obtain the image and the resolution provided by the different
sensors.
22 - What do Equity Hedge Funds Really do? Evidence in the
QE Period
Geum Il Bae, KAIST, 291, Daehak-ro, Yuseong-gu, Daejeon,
Korea, Republic of, gi_bae@kaist.ac.kr, Sun Young Park,
Woo Chang Kim
29 - Stochastic Optimization Methods for Nurse Staffing in
Inpatient Settings
Parisa Eimanzadeh, Wichita State University, 1845 Fairmount
Street, Wichita, KS, 67260, United States of America,
pxeimanzadeh@wichita.edu, Ehsan Salari
We examine why the hedge fund industry has experienced a slump during the
“Quantitative Easing (QE)” period. We analyze the risk-adjusted performances of
equity hedge funds in the pre-crisis, crisis, and QE periods. We show that the
disappeared alpha is the main reason for the inferior performance of hedge fund
industry these days, and reduction in exposure to systematic risks further explains
the underperformance of hedge funds in the QE period.
In this study, we use Queueing Theory and discrete-event simulation techniques
to determine nurse-staffing strategies that minimize staffing costs and ensure
timely delivery of nursing care to patients while accounting for the heterogeneity
in patients’ acuity and staff skill levels.
23 - NEOS Server: State-of-the-art Solvers for
Numerical Optimization
Rosemary T. Berger, University of Wisconsin - Madison, 330 N.
Orchard St., Madison, WI, 53715, United States of America,
rosemary.t.berger@gmail.com, Michael Ferris, Jeff Linderoth
30 - A Systems Dynamics Model for Flight Test
Knowledge Management
Roberto Follador, Mr, Institute for Advanced Studies - IEAv, Trevo
Coronel Av Jose A.A. Amarante, 01, Putim, Sao Jose dos Campos,
SP, 12228-001, Brazil, rcfollador@gmail.com
The NEOS Server is a free internet-based service for solving numerical
optimization problems. Hosted by WID at the University of Wisconsin in Madison,
the NEOS Server provides access to more than 60 state-of-the-art solvers in more
than a dozen optimization categories. Solvers run on distributed highperformance machines enabled by the HTCondor software. We describe recent
enhancements to the NEOS Server and highlight new interactive optimization
cases studies available on the NEOS Guide.
The research investigated how Knowledge Management (KM), in a Brazilian Air
Force (BAF) flight test environmen can be represented via a Systems Dynamics
Model. A documental research regarding the flight test environment KM was
done and a questionnaire was submitted to identify KM characteristics.
31 - A Supply Chain Network Equilibrium Model with Carbon
Capacity and Social Responsibility
Xiaoling Fu, School of Economics and Management,
Southeast University, Si Pai Lou 2#, Nanjing, 210096, China,
fufei1980@163.com, Lin Zhu, Xiangxiang Huang, Xiaogan Jiang
24 - Provable Submodular Function Minimization via
Wolfe’s Algorithm
Deeparnab Chakrabarty, Dr, Microsoft, 9 Lavelle Road,
Bangalore, India, deeparnab@gmail.com
This paper investigates a three-tier supply chain network equilibrium problem.
We first relate the decision makers’ social responsibility with transaction decisions
under the desired carbon capacity. Then we formulate the optimality of this
problem as a monotone variational inequality. Next, we propose a self adaptive
projection-based prediction—correction algorithm to solve the proposed model.
Finally, we report the numerical results and give some analysis on the equilibrium
solution.
Submodular function minimization (SFM) is an essential paradigm which appears
in many areas such as large scale learning and computer vision. The FujishigeWolfe Algorithm is agreed to be the fastest emprirical SFM algorithm. Despite its
good practical performance, very little is known about Wolfe’s minimum norm
algorithm theoretically. In this paper we give the first polynomial time
convergence analysis of Fujishige-Wolfe’s algorithm.
32 - How to Catch a Black Swan
David Gallop, Professor Of Program Management, Defense
Acquisition University, 6735 Surbiton Dr, Clifton, VA, 20124,
United States of America, davegallop@aol.com
25 - Stochastic PDE-constrained Optimization of Vibrations of a Plate
under a Piecewise-linear Current
Dmitry Chernikov, The University of Iowa, 1010 W Benton St.
#208F, Iowa City, IA, 52246, United States of America,
scher.de@gmail.com, Pavlo Krokhmal, Olesya Zhupanska
Projects are increasingly complex. We use risk-based management to address
complexity. Risk identification is the most important step in risk management
because risks that are unidentified are implicitly assumed. Group dynamics such
as silent dissent and group-think are weaknesses in team-based risk identification.
The PreMortem technique makes it safe for the team to address risks that may
otherwise go unidentified.
In this work a two-stage stochastic PDE-constrained optimization framework is
applied to the problem of vibration control of a thin composite plate in the
presence of electromagnetic field. The electric current is assumed to be of a
piecewise-linear form. We compute the gradient of the objective function using
adjoint numerical differentiation method. The value of the objective function is
calculated by solving the governing PDEs, and a black-box approach is used for
the minimization problem.
33 - Cost-effectiveness Analysis of Immunosuppression Therapy in
Primary Deceased Donor Renal Transplantation
Zahra Gharibi, SMU, 5507 Stonehenge Drive, Richardson, TX,
75082, United States of America, zgharibi@smu.edu,
Mehmet Ayvaci, Bekir Tanriover, Michael Hahsler
26 - Assessing Kernel-based Anomaly Detection Algorithms
Hyun-chang Cho, Seoul National University, Banpo-gu,
Seocho-dong, Seoul, Korea, Republic of, hccho@dm.snu.ac.kr,
Sungzoon Cho
The primary cure for patients with end stage renal disease (ESRD) is kidney
transplantation. In this study, we evaluate the cost-effectiveness of three common
immunosuppressive induction therapies, alemtuzumab, thymoglobulin, and
IL2RB as well as a no-induction strategy, from Medicare’s perspective. Using nonparametric bootstrapping method, we calculate the incremental cost-effectiveness
ratios for comparing the available strategies.
Anomaly detection is the process of finding items which do not comply with the
normal pattern of the data set. Although kernel-based approaches seem to be
promising for detecting anomalies, they have not been compared in a systematic
way. In this study, we generated numerous well-calibrated benchmark data set
and use them to evaluate the performance of various kernel-based anomaly
detection algorithms. The effect of kernel parameters will also be empirically
investigated.
313
POSTER SESSION
34 - The Effect of High Socioeconomic Inequalities on Public
Education Efficiency
Maria Cristina Gramani, Insper - Business Department,
Rua Quatá, 300, Vila Olímpia, São Paulo, SP, 04546042, Brazil,
mariacng@insper.edu.br
between a container terminal and different facilities.
A model is proposed to capture the full picture of educational efficiency in an
emerging country. Because of regional discrepancies, the model uses variables
related to education and to socioeconomic inequalities. The empirical results are
based on data from 5,129 Brazilian municipalities and the correlation factor
between the HDI-M and the educational efficiency score indicates that the HDI-M
index could not capture the discrepancies of a country with high levels of
socioeconomic inequality.
A motif-based evolutionary perspective is provided for infrastructure network
design. First, a multi-objective vulnerability-cost model is proposed to optimize
network structure. Secondly, an evolutionary algorithm is developed. Thirdly, a
network is tested by structure analysis, and motifs are traced during the
evolutionary process. Finally, Western States Power Grid is analyzed. Results have
revealed some principles in network design towards lower cascading vulnerability
and construction cost.
35 - Improving Scheduling and Control of the OHTC Controller in
Wafer Fab AMHS Systems
Shreya Gupta, Ms/phd Student And Graduate Research
Assisstant, University of Texas at Austin, Austin, TX, 78751,
United States of America, shreya.gupta@utexas.edu,
John Hasenbein
42 - Patient Reaction to Healthcare Data Breaches
Eric Johnson, Vanderbilt University, Owen School of
Management, Nashville, TN, United States of America,
eric.johnson@owen.vanderbilt.edu, Juhee Kwon
41 - Network Motif Analysis for an Infrastructure System
Against Vulnerability
Jing Jiang, PhD, Shanghai Jiao Tong University, Shanghai, China,
sjtujiangjing@163.com, Xiao Liu
We investigate consumer reaction to data breaches. Using a propensity score
matching technique, we analyze a matched sample of U.S. hospitals. We
investigate how breaches affect subsequent outpatient visits and admissions,
accounting for geographically-based competition. We find that the cumulative
effect of multiple breaches significantly decreases outpatient visits and admissions.
Automated Material Handling Systems (AMHS) in wafer fabs have complex
requirements. Thus, a larger number of AMHS vehicles are now required to pickup and transport these lots within the production facility. This has increased
vehicular traffic jams and the wait time for lots requiring pick-up and delivery.
Hence, to increase the system throughput, we present improved routing
algorithms for the over hoist transport control (OHTC) system.
43 - Enhancing Distribution Performance through Improved
Relationship Quality and Logistics Integration
Sung-tae Kim, Assistant Professor, SolBridge International School
of Business, 128 Uam-ro, Dong-gu, Daejeon, Korea, Republic of,
stkim1@solbridge.ac.kr, Moon-jung Yoo
36 - An Adaptive Large Neighborhood Search Heuristic for the
Inventory Routing Problem with Time Windows
Mina Hadianniasar, University of Arkansas, 901 N Pollard Street,
Arlington, VA, 22203, United States of America,
mhadiann@uark.edu, Ashlea Milburn
Prior research has argued that business relationship quality mediated by logistics
integration has shown positively related to distribution service performance.
Hence, firms attempt to achieve higher levels of logistics service and distribution
service performance through logistics integration. This study examines
relationship quality and logistics integration to understand how the two factors
are linked to distribution service performance.
This research models an integrated distribution and inventory control problem
(IRP) which is faced by a retail chain in the US. Currently, a direct shipping policy
with time window constraints is used for replenishing stores. This paper develops
an Adaptive Large Neighborhood Search Heuristic to determine the optimal
timing and magnitudes of deliveries to stores. The optimal plan considers direct
shipping policy as well as options combining deliveries for multiple stores into a
single route.
44 - Smart Logistics: Distributed Control of On-demand Green
Transportation Services
Seokgi Lee, Assistant Professor, University of Miami, 1251
Memorial Drive 281, Coral Gables, FL, 33146, United States of
America, sgl14@miami.edu, Yuncheol Kang, Vittaldas V. Prabhu
37 - Forecasting-based Truck Wait Time Reduction at Logistic Nodes
Alessandro Hill, Hamburg University of Technology,
Am Schwarzenbergcampus 4, Hamburg, Germany,
alessandro.hill@tuhh.de, Finn Meissner, Juergen Boese
We develop a strategic decision-making framework for on-demand delivery
services, considering both operational and environmental performance explained
by Just-In-Time delivery service, fuel consumption, and carbon emissions. The
optimal policies based on the Markov decision process are established to make
admission plans of delivery requests, and an integrated dynamic algorithm for
admission control and route scheduling is developed.
Truck wait times at logistic nodes such as container depots, packing facilities or
terminals cause delays in transport chains and traffic congestion. Truck companies
and nodes experience economical losses due to vehicle idle times and a lack of
planning reliability regarding routes, personnel or machinery, respectively. In this
work we present a flexible forecasting-based real worldapproach using artificial
neural networks to predict both, the truck wait times and the arrival rates at the
nodes.
45 - Extreme-point Search Heuristics Ffr Interval-flow Generalized
Network Problems
Angelika Leskovskaya, Southern Methodist University, 3145 Dyer
St., Suite 372, Dallas, TX, 75205, United States of America,
aleskovs@smu.edu, Richard Barr
38 - Impact of Overbooking in Appointment Scheduling of
Primary Care Services
Babak Hoseini, PhD Candidate, New Jersey Institute of
Technology, University Heights, Newark, NJ, 07102,
United States of America, bh77@njit.edu, Wenbo Cai
Interval-flow generalized networks are a new extension of the classic generalized
network formulation that adds a conditional lower bound constraint on the arcs.
An interval-pivoting heuristic that exploits the quasi-tree-forest basis structure to
explore extreme points is developed and computational testing is presented.
No-shows and late cancellations not only reduce the providers’ utilization, but
also results in long waiting time for other patients. Overbooking has the potential
to mitigate these negative impacts. However, excess overbooking may lead to
even longer waiting times for patients and prolonged working days for the care
team. We use a mathematical model to evaluates the benefit of overbooking and
develop a scheduling policy that reduces patients’ waiting time, and increase
provider’s utility.
46 - Hedge Fund Leverage Choice under
Time-inconsistent Preference
Bo Liu, University of Electronic Science and Technology of China,
No.2006, Xiyuan Ave,West Hi-Tech Zone, Chengdu, SC, 611731,
China, b.liu07@fulbrightmail.org
We show that time inconsistency preference discourages the manager from
underinvesting because of the high liquidation risk. The payment of incentive fees
may induce the irrational manager to be more aggressive and to overinvest.The
naive manager is more conservative than the sophisticated manager and prefers a
lower leverage level in normal times.Interestingly,investors are not sensitive to
the manager’s irrational investment behavior.
39 - Research on Combination of Container Yard Allocation and
Automatic Lifted Vehicle Path Optimization
Hongtao Hu, Shanghai Maritime University, Room101, No 96,
555 Guzong Road, Shanghai, China, hu.hongtao@foxmail.cm
This paper brings in a new type of automatic transport machinery—automatic
lifted vehicle which has the ability to lift container from the floor or put it down
on the floor. Meanwhile, a mixed integer programming model is established to
ensure that all the containers handled as far as possible in the time window. The
model also considers the problem of allocating blocks to discharge containers and
optimizing path of automatic lifted vehicle.
47 - New Assay Implementation Planning at Clinical Laboratory
Wei Liu, Industrial Engineer, MD Anderson Cancer Center, 8515
Fannin St, Houston, TX, 77054, United States of America,
wliu8@mdanderson.org, Cindy Lewing, Bedia Barkoh,
Pramod Mehta, Mark Routbort, Humin Lu, Justin Villarreal,
Raja Luthra, Keyur Pravincha Patel, Geeta S Mantha,
Mylene Bole, Haobo Yang, David Garcia, Zou Zhuang
40 - Shipping Commodities Between a Container Terminal and
Different Destination Zones using Heavy Trucks
Mazen Hussein, Assistant Professor, University of WisconsinPlatteville, Platteville, WI, 53818, United States of America,
husseinm@uwplatt.edu
Implementation of a new complex laboratory assay at our high-volume and highcomplexity clinical Molecular Diagnostics Laboratory was facilitated by
application of multiple engineering approaches including workflow assessment,
historical volume-based demand prediction, IT solution, and resource allocation.
The new assay implementation is expected to be successful with minimal
workflow interruptions, no patient care interruptions, low implementation cost
and optimal resource utilization.
The cost model for shipping commodities by truck developed by Hussein M.
(2010) is extended to consider the impact of tollway polices on truck route
selection for shipping containers of specific commodity groups near a container
terminal. A path-finding model is built for this purpose. The values of time were
used to mimic the truck’s choices to ship containers of different commodities
314
48 - Big Data Analytics for Singapore Public Train System
Nang Laik Ma, Senior Lecturer, SIM University, 461 Clementi
Road, Singapore, 599491, Singapore, nlma@unisim.edu.sg,
Beng Yee Wong
POSTER SESSION
evaluated according to engineering, environmental, and economic (3E) criteria
using a cradle-to-gate life-cycle approach.
55 - Analysis on the Effect of Energy Efficient Technologies in
Industry Sector using Times Model
Sang Yong Park, Senior Researcher, Korea Institute of Energy
Research, Yuseong-gu, Gajeong-ro 152, Daejeon, 305-343, Korea,
Republic of, gspeed@kier.re.kr, Jong Chul Hong, Nyunbae Park,
Boyeong Yun
This paper focus on capacity planning of the Singapore public transport system.
We analyse the commuters’ travelling patterns from historical transactions data.
Secondly, by simulating train schedule and capacity constraints, the model
mimicked the real-world situations to generate the waiting time for each
commuter. Finally, a web-based visualization tool is created to provide the
average waiting time for the next train at the station to enhance the commuter’s
experience.
49 - Hospital Residents Problem: A Survey Including a New Variant
Kaitlyn Manley, College of Charleston, 66 George St, Charleston,
SC, 29424, United States of America, manleykm@g.cofc.edu,
Tyler Perini, Amy Langville
The South Korea established energy policy which is focusing demand
management rather than energy supply to secure a stable energy supply and to
cope with climate change efficiently through 2nd national energy basic plan in
2014. This research developed energy system model which can analyze the effect
of energy efficient technologies on demand management based on TIMES(The
Integrated MARKAL-EFOM System) model and conducted case study on industry
sector in Korea.
We survey several variations of the Stable Matching Problem, including the
Hospital Residents Problem used to assign American medical residents to
hospitals. We also present a new variation of the stable matching problem that
uses an binary integer linear program to determine the minimum number of
interviews that hospitals should conduct in order to still maximize the number of
residents assigned.
56 - The Humility Project: NMF and Other Matrix Factorizations for
Textual Analysis
Tyler Perini, Student, College of Charleston, 66 George St,
Charleston, SC, 29424, United States of America,
perinita@g.cofc.edu, Amy Langville
50 - Exploring the Multi-objective Portfolio Tradespace
Simon Miller, Graduate Student, Penn State Applied Research
Lab, P.O. Box 30, State College, PA, 16804, United States of
America, ses224@arl.psu.edu, Gary Stump, Sara Lego,
Michael Yukish
This is one of the first studies on the use of matrix decompositions as the primary
engine for describing and predicting psychological characteristics in a corpus of
language data. With text parsing tools, large written samples are parsed into a
sparse matrix. A low-rank matrix factorization of a weighted version of this
matrix is then used to determine which documents are humble and which are not
humble. Three factorizations, the SVD, NMF, and weighted NMF, are compared.
Faced with strategic choices, senior decision makers must often make trades to
meet competing requirements. In collaboration with the U.S. Army, ARL has
developed tools and methods to treat large scale, multi-objective optimization
problems for binary portfolios with dynamic constraints. Methodology and
implementation schema for real-world cases are presented, highlighting the
ability to balance a combinatorial explosion of parameters in complex trades
spaces with the need for timely decisions.
57 - Distributed Online Modified Greedy Algorithm for Networked
Storage Operation under Uncertainty
Junjie Qin, PhD Candidate, Stanford University,
126 Blackwelder Ct, 1004, Stanford, CA, 94305,
United States of America, jqin@stanford.edu
The optimal control of energy storage networks in stochastic environments is an
important open problem. This paper provides an efficient algorithm to solve this
problem with performance guarantees. A sub-optimality bound for the algorithm
is derived which can be minimized by solving a semidefinite program. Distributed
implementation is derived based on alternating method of multipliers. Numerical
examples verify our theoretical performance bounds and demonstrate the
scalability of the algorithm.
51 - Experimental Designs for Metal Detectors at Large Venues
Christie Nelson, Rutgers University, CCICADA, 4th Floor, CoRE
building, 96 Frelinghuysen Rd, Piscataway, NJ, 08854, United
States of America, christie.l.nelson.phd@gmail.com, Paul Kantor,
John Edman, Vijay Chaudhary
Walk-through metal detectors (WTMDs) are being used increasingly more as a
security measure at large events, particularly at stadiums. Currently, WTMDs are
tested using a robotic tester which tests metallic objects at level heights by sending
them straight through at a constant speed. However, this is not a proper
representation of how a person would enter the WTMD. We will show that the
way a person walks through the WTMD impacts detection rate through our
experimental results.
58 - Should Retailers Adopt 3d Printing?
Sharareh Rajaei Dehkordi, PhD Candidate, New Jersey Institute
of Technology, University Heights, Newark, NJ, 07102, United
States of America, sr552@njit.edu, Wenbo Cai
Should retailers provide 3D printing services in addition to the traditional off-theshelf product? We answer the question by examining retailers’ optimal joint
decisions on his inventory management policy and pricing scheme while
considering consumers’ heterogeneous preferences for self-designed, 3D printed
products vs. off-the-shelf products. We use a multi-server queue with limited
capacity to capture customers’ production selection process and its impact on the
retailer’s expected profit.
52 - Leading Metrics for Progress Measurement and Performance
Assessment in Construction Projects
Resulali Orgut, Graduate Research Assistant, North Carolina State
University, Dept. of Civil, Cons. and Env. Eng., 2510 Stinson Dr.,
222 Mann Hall, Raleigh, NC, 27695, United States of America,
reorgut@ncsu.edu, Jin Zhu, Mostafa Batouli, Ali Mostafavi,
Edward Jaselskis
59 - Stochastic Network Design with
Decision-dependent Uncertainties
Nathaniel Richmond, University of Iowa, 14 MacLean Hall, Iowa
City, IA, 52242, United States of America,
nathaniel-richmond@uiowa.edu
Progress measurement and performance assessment are critical to the
management of construction projects. We perform statistical analyses to highlight
key indicators for successful construction project controls by using data collected
through an online survey from 27 companies. We analyze core metrics commonly
used in the construction industry to develop guidelines for improving their
reliability and recommend practices for interpreting metrics and indicators.
Little research has been conducted on stochastic network design problems in
which the probability distribution of future random events is affected by prior
actions. However, such problems are ubiquitous and important. For example,
planned reinforcements of a power network directly influence which nodes are
more likely to fail. We present a stochastic two-stage programming model with
decision-dependent uncertainties, discussing solution methods for the associated
unique computational challenges.
53 - Toward Consistent and Efficient Anomaly Detection in
Hyperspectral Imagery
Todd Paciencia, USAF, AF/A9, Pentagon, Washington D.C.,
United States of America, todd.j.paciencia.mil@mail.mil
60 - Scheduling Part-time Employees with Interactive Optimization
Robert Rose, President, Optimal Decisions LLC, 4 Kirby Lane,
Franklin Park, NJ, 08823, United States of America,
robertl.rose@verizon.net
This research will showcase development of an approach to making an
unsupervised anomaly detector for Hyperspectral Imagery (HSI). The algorithm is
developed to be robust to different image scenes, different sensors, and noisy
spectral bands. Specifically, fusion of spectral, spatial, and Signal-to-Noise
information is used, in combination with a factor analysis approach, to identify
anomalies. The algorithm is shown to be desirable when compared to current
state-of-the-art techniques.
Many employee scheduling problems are very challenging: they are hard
combinatorial optimization problems that contain multiple objectives and ‘soft’
constraints. Such problems do not lend themselves to a pure optimization
approach. A ‘Human Centered’ approach, will be described: an initial schedule is
generated analytically through a series of heuristic procedures, and a final
schedule is produced using an interactive graphics module. A prototype
scheduling program will be demonstrated.
54 - Comprehensive Performance Evaluation of High-gravity
Carbonation Process in the Steelmaking Industry
Shu-Yuan Pan, National Taiwan University, No 71 Chou-Shan
Rd., Taipei, 10673, Taiwan - ROC, d00541004@ntu.edu.tw,
Pen-chi Chiang
An integrated portfolio of multi-waste treatment (steelmaking slag and
wastewater) combined with CO2 capture in the steelmaking industry can be
achieved by the high-gravity carbonation (i.e., HiGCarb) process using a rotating
packed bed (RPB). In this study, the HiGCarb process was comprehensively
315
POSTER SESSION
61 - Fast, Provable Algorithms for Isotonic Regression in All
Lp-norms (to Appear At Nips 2015)
Sushant Sachdeva, Postdoctoral Associate, Yale University, Yale
Institute of Network Sciences, P O Box 208263, New Haven, CT,
06520, United States of America, sachdevasushant@gmail.com,
Rasmus Kyng, Anup Rao
67 - An Optimization Approach to Warehouse Line Striping
Sudharshana Srinivasan, Research Scientist, Altria Client
Services, 601 E. Jackson St., Richmond, VA, 23219, United States
of America, sudharshana.srinivasan@altria.com, David Kane
We present a mixed-integer programming model to optimize product storage at
an industrial warehouse, while adhering to safety standards stipulated by the
county and the business. The model is applied to a tobacco warehouse and the
results provide a storage solution comparable to current practice with an
improved operational efficiency. The model recommends more walking aisles to
facilitate increased spacing and airflow around the product; both of which are
valued business objectives.
Given a directed acyclic graph G, and values y on the vertices, the Isotonic
Regression of y is a vector x that respects the partial ordering given by G, and
minimizes ||x - y||, for a given norm. We present improved algorithms for Isotonic
Regression for all weighted Lp norms, with rigorous performance guarantees. Our
algorithms combine interior point methods with provable fast solvers for the
associated linear systems. The algorithms are practical and lend themselves to fast
implementations.
68 - Testing the Applicability of Genetic Algorithms for
Simulation-based Healthcare Optimization
Cory Stasko, MIT, 4 Garden Court, Apt. #4, Cambridge, MA,
02138, United States of America, cstasko@mit.edu
62 - Mathematical Modelling and Analysis of New Zealand
Legislation Network
Neda Sakhaee, University of Auckland, 38 Princes Street,
Auckland, New Zealand, nsak206@aucklanduni.ac.nz
In 2015 the concept of Legislation Network is proposed as a mathematical tool for
studying the current and future status of the legislation system in European
Union. Unlike perhaps the relations between documents are at least as important
as the content. This type of network has some novel features which make it an
excellent test case for new network science tools.
We apply genetic algorithms to three distinct cases of highly non-linear healthcare
optimization problems. In the first problem, OR schedules are designed to
minimize downstream bottlenecks. The second problem involves network
management for an accountable care organization. The third problem involves
promoting the spread of ideas among connected professionals. In each case, the
objective (fitness) function is the output of a simulation, and a brute force
solution search is not feasible.
63 - Deadhead Selection Strategies for Crew Recovery
Sujeevraja Sanjeevi, Senior Operations Research, Sabre Holdings,
3150 Sabre Drive, Southlake, TX, 76092, United States of
America, sujeevraja.sanjeevi@sabre.com, Chunhua Gao,
Helder Inacio
69 - How Do We Capture the Potential Risk of Intravenous Drug
Infusion using Alert Data?
Wan-ting Su, Graduate Student, Purdue University, 3376
Peppermill Drive, West Lafayette, IN, 47906, United States of
America, su33@purdue.edu, Poching Delaurentis, Mark Lehto
Crew recovery is the problem of minimizing the impact of a disruption to an
airline by getting disrupted crews back on plan while minimizing incurred costs.
Deadheads are flights that transport crew members as passengers and are a critical
part of crew recovery. Consideration of all available deadheads to recovery makes
problem sizes prohibitive. We present a few deadhead selection strategies that
significantly improve solution quality without impacting run-time for real-world
scenarios.
The use of smart infusion pumps is one such mechanism in ensuring the safety of
medication infusions in clinical settings. We aim to utilize the data of different
alert types from the Infusion Pump Informatics system to capture averted or
potential medication errors and define and determine the overall risk of potential
harm within a certain period of time in a medical facility. Our analysis can be
used as a measure in improving intravenous medication safety and infusion drugdelivery process.
64 - Talk is Cheap - Action is Expensive
Simone Schmid, University of Chemnitz, Huebschmanntrasse 24,
Chemnitz, 09112, Germany, simone.schmid@wirtschaft.tuchemnitz.de, Peter Pawlowsky
70 - Developing Freeway Demand Estimation Alternatives with Mixed
Integer Linear Programming
Joseph Trask, North Carolina State University, 909 Capability
Drive, Raleigh, NC, United States of America, jltrask@ncsu.edu,
Behzad Aghdashi, John Baugh, Nagui Rouphail
Adequate response to uncertain and unpredictable environmental changes
requires innovative, agile, and adaptive team competencies. We use an
interdisciplinary approach to assess and evaluate team competencies. From theory
and previous research we derive indicators and test these by training teams
accordingly. Experimental groups were given theoretical and practical trainings
with regard to these team competencies. Control groups acted as usual. A
succeeding standardized simulation in a high fidelity simulation environment
showed significant effects with regard to team performance. From these results
we propose behavioral markers for team competencies that can be used to assess
team performance in critical situations.
This research presents a Mixed Integer Linear Programming optimization model
for traffic demand estimation based on the methodology developed in the
Highway Capacity Manual (HCM). Due to a lack of uniqueness for solution
demand sets, a Modeling to Generate Alternatives (MGA) approach is developed
to investigate the wide ranges of optimal solution sets. These solution sets can be
compared through their effects on intermediate performance measures and
sensitivity analysis.
71 - Batch Testing of a Series System
Tonguç ‹nlöyurt, Sabanci University, Orhanli, Tuzla, Istanbul,
Turkey, tonguc@sabanciuniv.edu, Ozgur Ozluk, Rebi Daldal,
Baris Selcuk, Zahed Shahmoradi
65 - Euro/Roadef Challenge
Tejinder Singh, Air Liquide, 12800 W. Little York Rd, Houston,
TX, United States of America, tejinder.singh@airliquide.com,
Jean Andre, Michele Quattrone, Rodrigue Fokouop,
Jeffrey Arbogast
We consider the problem of determining the correct value of an AND function
when it is costly to learn the values of its variables, with the minimum expected
cost. We refer to a subset of variables whose values can be learnt at the same time
a meta-test. The cost of learning the values of the variables in a meta-test includes
a fixed cost plus the costs of the tests in the meta-test.
The French OR Society (ROADEF) along with EURO, periodically organizes an
OR challenge dedicated to industrial applications. This year, Air Liquide proposes
the challenge problem concerning an IRP for the bulk distribution of liquefied
gases. The challenge is open to everyone and will be presented during the EURO
2015 in Glasgow, Scotland in July 2015 and the results will be announced at
EURO 2016 in Poznan, Poland. Prizes totaling 20,000 Euros will be awarded to
the best teams.
72 - Shape-preserving L1 Spline Fits: Calculation and Capability
Ziteng Wang, Department of Industrial and Systems Engineering,
North Carolina State University, 3120 Walnut Creek Parkway,
Apt. E, Raleigh, NC, 27606, United States of America,
zwang23@ncsu.edu, Tiantian Nie, Shu-cherng Fang
66 - Usability Evaluation of a Mobile App to Reduce Congestive
Heart Failure (CHF) Readmissions
Minal Singhee, Master’s Student, H. John Heinz III College,
Carnegie Mellon University, 5000 Forbes Avenue, Pittsburgh, PA,
15213, United States of America, msinghee@andrew.cmu.edu,
Daniel Gartner, Rema Padman, Jina Lee, Sriram Iyengar
L1 spline fits have been developed over the past decades to approximate multiscale data and have been shown to preserve shapes well. Local approaches are
designed for efficient calculation. Quantitative measures are proposed to evaluate
the shape-preserving capability of different types of L1 spline fits.
73 - Evolutionary Optimization Tools of Nanostructures for Solar Cells
Baomin Wang, University of Pittsburgh, 1048 Benedum Hall 3700
O’Hara Street, Pittsburgh, PA, 15261, United States of America,
baw57@pitt.edu
Congestive Heart Failure (CHF) is a major chronic condition affecting more than 5
million people in the US. CHF readmissions is one of the major contributors to the
burgeoning healthcare cost. In our study, we evaluate factors associated with the
usability and acceptability of a mobile application intended to reduce CHF
readmissions. A fixed protocol was developed which included Think Aloud
Protocol, Quiz and a Questionnaire. Qualitative and quantitative analyses were
performed.
Simulation plays a significant role in optimizing solar cell efficiency. Current used
optimization is exhaust search, only feasible for small group of parameters, 3 or 4.
But for 7 or 8, it takes months. In this work, we integrate genetic algorithm with
FDTD methods to optimize the nanostructure. This evolutionary method can
decrease the simulation time to 1/6 of original time. This work demonstrates the
ability of genetic algorithm technique to quickly search through a large parameter
space.
316
TC01
74 - Surgical Operations Scheduling with Machine Eligibility and
Resource Constraint
Shan Wang, Shanghai Jiaotong Univerisity, 704 West 180th
Street, First Floor, #4, New York City, NY, 10033, United States of
America, wangshan_731@sjtu.edu.cn, Guohua Wan, Huiqiao Su
driving forces of the underlying community structures and understand the
relation to cost and utilization.
We study a problem in surgical operations scheduling and model it as a resourceconstrained machine scheduling problem with eligibility restriction to minimize
the makespan. By decomposing it into two subproblems, we develop effective
heuristic algorithms to solve the problem. We test the proposed algorithms on
randomly generated instances as well as real data set from a large hospital. The
numerical results show the effectiveness and potential practical value of the
models and the algorithms.
Saudi Aramco Engineering Services developed a systematic model to calculate
shell and tube heat exchangers efficiency in real time using data analytics
technique. Monitoring the heat exchanger efficiency in real time supports the
decisions making to plan for turndown time by detecting the failure time
proactively. Moreover, it provides the field engineers continuous monitoring to
energy consumer and highlights the wasted energy locations, quantity and cost.
75 - Optimizing System-Level Preventative Maintenance Cost of
Multistate Series-Parallel Systems
Sallamar Worrell, The George Washington University, 3117
Icehouse Place, Bryans Road, MD, 20616, United States of
America, skaw7@gwu.edu, James Moreland Jr., Steve Doskey
82 - A Decision-analytic Approach to Inferring Preference from
Choice Behavior
Matthew D. Wood, Research Psychologist, US Army Engineer
Research & Development Center, Concord, MA, United States of
America, Matthew.D.Wood@usace.army.mil, Matthew Bates,
Danielle M. Beeler, Jeffrey M. Keisler
Maintenance costs for complex systems are often overpaid due to the lack of
maintenance harmony between the individual subsystems. The research in this
study proposes the use of a new meta-heuristic optimization method, the Grey
Wolf Optimizer algorithm developed by Mirjalili and Lewis, to identify the
optimal system-level maintenance strategy for multistate series-parallel systems
and aims to produce better results than the methods previously applied to this
problem in published literature.
Decision making in resource management requires consideration of multiple
conflicting objectives. Traditional elicitation methods (e.g., swing weighting) work
well with clearly defined problem spaces and effects between decisions and
consequences. In more complex domains, an inferential approach is needed to
estimate preference using system feedback over samples with subject-matter
experts. We describe a decision analysis game with an environmental case study
problem.
76 - Predicting Digital Currency Price from Social and
Traditional Media
Peng Xie, Georgia Institute of Technology, Room 907,
100 10th Street, Atlanta, GA, 30309, United States of America,
peng.xie@scheller.gatech.edu
83 - Artificial Variability and Case Mix in Relation to Patient Flow
at a Hospital Outpatient Clinic
Monique Bakker, City University of Hong Kong, Hong Kong,
999077, Hong Kong-PRC, moniquebakker121@gmail.com
81 - Real-time Heat Exchanger Efficiency Monitoring
Yousif Abualsoud, Saudi Aramco, Al-Midra, Dhahran,
Saudi Arabia, yousif.abualsoud@aramco.com
We aim to improve the patient flow through first, second, and nth visits to a
specialist outpatient clinic and elective surgery. We investigate variability in key
resource availability (i.e. the specialists) in relation to case mix decisions, subspecialization restrictions, and resource allocation: how is indirect wait time (or
access time). We use Discrete Event Simulation to schedule patients under a
broad and realistic set of rules and restrictions to compare alternate scenarios.
Using daily Bitcoin price data and Bitcoin Forum discussion, we try to understand
if social media can affect Bitcoin price and how long does it take. We use the
percentage of negative words as the measure of the article sentiment. The results
show that social media can affect price. However, for information sources focusing
on speculation, the effects on prices are immediate. In contrast, information
concerning fundamentals impacts prices in a longer holding period.
77 - The Dimensions of Supervision-subordinate Relationship
Liuqing Yang, Xi`an Jiaotong University, No.28, Xianning West
Road, Xi’an, China, ylq2011@stu.xjtu.edu.cn, Qiaozhuan Liang
Subordinate can form distinguishable social exchange relationships with their
immediate supervisor.The exchange resources differ between organizational and
personal. Based on the taxonomy, social exchange relationship can development
a construct of four dimensions: work, flatter, private, selfless. Various dimension
have different impact on performance as a mediator or intervening variable. The
construct can explain the puzzle of relationship between social exchange
relationship and performance.
■ TC01
Logistics and Operations Research
Sponsored Session
78 - Counterfeits, Anti-counterfeit Technology and
Monitoring Strategy
Shiqing Yao, The Chinese University of Hong Kong, 9/F, Cheng
Yu Tung Building, No.12 Chak Cheung Street, Shatin, N.T., Hong
Kong, Hong Kong - PRC, sqyaocuhk@gmail.com, Kaijie Zhu
Chair: Andrew Hall, COL, U.S. Army, 4760 40th St N, Arlington, Va,
United States of America, AndrewOscarH@aol.com
1 - Scheduling the Transition of the C-130 Aircraft Fleet to a New
Maintenance Process
Melissa Bowers, mrbowers@utk.edu, Bogdan Bichescu,
Kenneth Gilbert, Anurag Agarwal, Doug Keene
We consider an authentic company that sells its products to a customer market. In
the same market, a counterfeiter may sell its low-quality counterfeits. The
authentic company can put effort into developing an anti-counterfeit technology
to distinguish its products from counterfeits while the authority can monitor the
market and outlaw counterfeiting localities. We derive the company’s optimal
decision on its anti-counterfeit effort and highlight its difference from
conventional wisdom.
The US Air Force is tasked to increase aircraft availability through a new
initiative, the High Velocity Maintenance program, which seeks to increase
availability of aircraft to the US military through a reduction in the time an
aircraft spends on the ground undergoing planned maintenance. The transition of
a fleet of aircraft from the standard process to shorter HVM cycles is a complex
task that requires careful planning. A mixed-integer programming approach is
used to generate a schedule which maintains a constant annual workload and
yields lower flow times and work-in-process levels.
79 - Optimization of Area Traffic Control: A Binary Mixed Integer
Linear Programming Approach
Zhao Zhang, Research Assistant, Tsinghua University, Room 615,
Shunde Building, Haidian District, Beijing, 100084, China,
zzaxx@tsinghua.edu.cn
2 - Learning and Bayesian Updating in Made-to-order
(MTO) Production
Keith Womer, Professor, University of Missouri - St. Louis, One
University Blvd, St. Louis, MO, 63121, United States of America,
womerk@umsl.edu, Jeffrey Camm, Haitao Li, Colin Osterman,
Rajesh Radhakrishnan
This paper proposes a model aims at optimizing area traffic control. We use
network total delay as the objective in the model. In this research, cell
transmission model is used to discretize research time into many intervals and
signal coordination, lane settings, phase, start of green and green split can be
optimized simultaneously. The model is linear in nature and can be solved by
standard branch and bound algorithm.
The case of production planning for made-to-order (MTO) manufacturing. We
minimize expected discounted cost by controlling production rate. A dynamic and
adaptive approach to estimate the effects of learning and to optimize next period
production is developed. This approach offers a closed-loop solution through
stochastic dynamic programming. The approach is illustrated using data from the
Black Hawk Helicopter Program.
80 - Identify Naturally Occurring Healthcare Provider Referral
Networks for Diabetic Patients
Yuchen Zheng, PhD Student, Georgia Institute of Technology,
Atlanta, GA, 30032, United States of America,
richardzyc@gatech.edu, Kun Lin, Jeremy Pickreign,
Thomas White, Gigi Yuen
The pioneer Accountable Care Organizations (ACOs), where doctors voluntarily
form groups to deliver coordinated quality care, saved Medicare $400 million in
two years. To improve ACOs design, we utilize past care patterns and adopt
modularity maximization to detect pre-existing referral networks for diabetic
patients, within which doctors share patients frequently. We further identify the
317
TC02
3 - Retail Inventory Optimization for the U.S. Navy
Javier Salmeron, Naval Postgraduate School, Operations Research
Dept., GL-214, Monterey, CA, 93943, United States of America,
jsalmero@nps.edu, Emily Craparo
■ TC03
We present a mixed-integer linear model to guide retail (site-based demand level)
inventory decisions for the Naval Supply Systems Command (NAVSUP), Weapons
Systems Support. An (s,Q) model optimizes reorder points and order quantities
for thousands of items in order to minimize deviations from target fill rates, with
demand uncertainty, budget and order quantity constraints. We compare branchand-bound with Lagrangian relaxation solutions, and our results with those from
other NAVSUP tools.
Contributed Session
Inventory Management I
Chair: Jun-yeon Lee, California State University, Northridge, 18111
Nordhoff St, Northridge, CA, 91330-8378, United States of America,
junyeon.lee@csun.edu
1 - Forecasting of Demand Tail Distribution for Inventory Optimzation
Antony Joseph, Staff Data Scientist, Walmart Labs,
1001 National Avenue, Apt. 107, San Bruno, CA, 94066,
United States of America, AJoseph0@walmartlabs.com
■ TC02
We discuss a novel technique for forecasting demand tail distribution for items in
the Walmart e-commerce inventory. The methodology first estimates the
quantiles of the demand distribution, followed by fitting a parametric distribution
using a suitable metric. The method is seen to be robust to high observed
variability in demand. Performance of the proposed approach is assessed using
Supply Chain metrics such as Weeks of Supply and Met Demand.
Network Applications in Homeland Security
Invited Session
2 - Strategic Safety Stocks under Guaranteed Service and
Constrained Service Models
Ton De Kok, TU Eindhoven, Paviljoen E.04, P.O. Box 513,
Eindhoven, Netherlands, a.g.d.kok@tue.nl
Chair: Thomas Sharkey, Rensselaer Polytechnic Institute, 110 8th
Street, Troy, NY, 12180, United States of America, sharkt@rpi.edu
1 - Novel Bilevel Programming Approaches for Interdicting
Multi-tiered Illegal Supply Chains
N. Orkun Baycik, Rensselaer Polytechnic Institute, 110 8th Street,
Center for Industrial Innovation, Troy, NY, 12180, United States
of America, baycin@rpi.edu, Thomas Sharkey, Chase Rainwater
In this presentation we discuss our findings on a set of real-life supply chains
concerning the positioning of safety stocks in the supply chain. We compare the
results from models under the guaranteed service assumption and under the
constrained service assumption. The latter assumption follows the classical Clark
and Scarf model for serial multi-echelon systems. Furthermore, we discuss some
implications of the guaranteed service assumption in case of a bounded demand
formulation.
We study a resource allocation problem faced by law enforcement in arresting
criminals in the drug flow and information flow networks. The objective of law
enforcement is to minimize the maximum amount of drugs reaching end users.
There exist interdependencies between the networks which leads to a network
interdiction problem with a discrete inner problem. We apply a novel dual-based
reformulation technique to solve an equivalent single-level problem and present
computational results.
3 - Scarcity Effect on Dual-channel Supply Chain
Baoshan Liu, PhD Student, Huazhong University of Science and
Technology, School of Management, 1037 Luoyu Road, Wuhan,
China, liubaooshan@qq.com, Shihua Ma
2 - Bi-level Stochastic Network Interdiction Model for Deployment of
Cyber-security Countermeasures
Apurba Nandi, Mississippi State University, P.O. Box 9542,
Mississippi State University, Mississippi State, MS, 39762,
United States of America, akn77@msstate.edu, Hugh Medal
We consider manufacturer’s dual-channel sale system where customers get the
product the direct channel with limited quantity. The limited quantity of the
direct channel releases a scarcity message that consumers will increase their
purchasing desire. Both the manufacturer and the retailer choose their own
decision variable to maximize their respective profits considering scarcity effect.
We model the problem using Stackelberg games and try to find the best solution
from different perspectives.
We study how best to deploy cyber-security countermeasures to protect a cybernetwork against attacks. We propose a bi-level stochastic network interdiction
model capturing the interaction between the attacker and the defender as a
sequential stackelberg game played on an attack graph. We consider that the
attacker’s knowledge about the topology of the attack graph, and the attacker’s
and the defender’s knowledge about each other’s actions are uncertain. We
develop algorithm to solve the model.
4 - An Optimal Inventory Replenishment Considering Product
Life Cycle
Shunichi Ohmori, Assistant Professor, Waseda University, Room
51-15-05, Okubo 3-4-1, Shinjuku, Tokyo, 169-8555, Japan,
ohmori0406@gmail.com, Kazuho Yoshimoto
3 - Vitality Measures for Multigraphs with Applications to
Communications Networks
Sarah Nurre, Assistant Professor, Air Force Institute of
Technology, 2950 Hobson Way, WPAFB, OH, 45433, United
States of America, Sarah.Nurre@afit.edu, Christopher Hergenreter
We consider a problem of determining initial and replenishment order quantities
that minimize the cost of lost sales, inventory holding cost, fixed ordering cost,
and obsolete inventory subject to stochastic demands. We model this problem as a
multi-stage problem where the demands and prices of products lie in some
uncertainty set depending on their life cycle. We develop a dynamic programming
method to solve this problem.
We consider the problem of determining the most vital arcs to protect within a
multigraph, such as a communications network. Traditional vitality measures are
insufficient as they often examine single arc failures which have no impact on
multigraphs due to parallel arcs between pairs of nodes. Herein, we propose and
examine set based vitality measures for multigraphs. We perform and present the
results of promisting computational results multi-mode military communications
networks.
5 - Vendor-managed Inventory with a Time-dependent
Stockout Penalty
Jun-yeon Lee, California State University, Northridge, CA,
18111 Nordhoff St, Northridge, CA, 91330-8378,
United States of America, junyeon.lee@csun.edu
We examine the problem of designing a vendor-managed inventory (VMI)
contract with a time-dependent stockout penalty in a stochastic (Q, r) inventory
system, in which the supplier is charged a stockout penalty that depends on the
length of the time during which stockouts occur at the customer. The customer
chooses the stockout penalty and offers the VMI contract to the supplier, and the
supplier can accept or reject the contract. We examine the optimal behaviors of
the two parties.
4 - Efficient Resilience Optimization of Interdependent Networks
Andres Gonzalez, Rice University, 6100 Main Street, MS-318,
Ryon 203, Houston, TX, 77005, United States of America,
andres.gonzalez@rice.edu, Leonardo Dueñas-osorio,
Andres Medaglia, Mauricio Sánchez-silva
MIP models of the Interdependent Network Design Problem (INDP) have proved
to be effective tools to study and improve the resilience of systems of
interdependent infrastructure networks. Nevertheless, these MIP models have
limited scalability for large realistic systems. In this work, we present an enhanced
methodology to optimize the resilience of interdependent networks, based on the
joint use of decomposition techniques and efficient reformulations of the INDP
models.
318
■ TC04
TC06
In this paper, we undertake a content analysis of the news feed regarding big data
analytics. We develop a corpus of related news items. We code the corpus and
prepare it for analysis by the Natural Language Processing (NLP) algorithms. Our
purpose is to conduct a systematic analysis of the news contents to determine the
primary themes being projected by the proponents of big data. We will explore
ways to identify real value content that may help to draw meaningful inferences.
Social Media Analytics Best Papers Finalist
Competition
Invited Session
2 - Modeling Message Dissemination in Social Media for
Emergency Communication
Justin Yates, Professor, Francis Marion University,
4822 E Palmetto St, Florence, SC, United States of America,
jyates@fmarion.edu, Xin Ma
Chair: Theodore Allen, Associate Professor, The Ohio State University,
1971 Neil Avenue, 210 Baker Systems, Columbus, OH, 43221, United
States of America, allen.515@osu.edu
1 - Social Media Analytics: The Effectiveness of Marketing Strategies
in Online Social Media
Vilma Todri, PhD Candidate In Information Systems, NYU, 44 W
4th St, KMC Room 8-181, New York, NY, 10012, United States of
America, vtodri@stern.nyu.edu, Panagiotis Adamopoulos
We introduce the Single-network Multi-message Social Media Message
Dissemination Problem (SM-SMMDP) as a discrete optimization model to
examine message dissemination in social media and to explore message
dissemination strategies for government and non-government emergency
management agencies. We implement a detailed computational experiment on
representative small-scale networks and discuss implications for real-world
application.
This paper studies a novel social media venture and seeks to understand the
effectiveness of marketing strategies in social media platforms by evaluating their
impact on participating brands. We use a real-world data set and employ a
promising research approach combining econometric with predictive modeling
techniques.
3 - Analyzing and Predicting Threatening Language in
#gamergate Tweets
Cheyanne Baird, Senior Linguistic Specialist, SAS Institute Inc.,
Boston Regional Office, Prudential Tower, 800 Boylston St.,
Suite 2200, Boston, MA, 02199, United States of America,
Cheyanne.Baird@sas.com, Michael Wallis, Praveen Lakkaraju
2 - Predicting Crowd Behavior with Big Public Data
Nathan Kallus, MIT, 77 Massachusetts Ave., E40-149, Cambridge,
MA, 02139, United States of America, kallus@mit.edu
Aggressive language in online communities can thrive, escalate, and signal
palpable threat. This is apparent in #gamergate tweets. Using SAS Text Analytics,
we will show how language transitions from negative sentiment to threatening
language in #gamergate tweets, building a model to predict probable threat in
social media.
We present efforts to predict the occurrence, specific timeframe, and location of
crowd actions before they occur based on public data collected from over 300,000
open content web sources in 7 languages, from all over the world, ranging from
mainstream news to government publications to blogs and social media.
3 - Participation vs. Effectiveness of Paid Endorsers in Social
Advertising Campaigns: A Field Experiment
Jing Peng, The Wharton School, University of Pennsylvania,
3730 Walnut Street Suite 500, Philadelphia, PA, 19104,
United States of America, jingpeng@wharton.upenn.edu,
Christophe Van den Bulte
■ TC06
Dynamics and Information in Commodity Markets
Sponsored Session
We investigate the participation and effectiveness of paid endorsers in viral-forhire social advertising. Specifically, we investigate (i) how financial incentives
affect potential endorsers’ participation and effectiveness in generating online
engagements (likes, comments, and retweets), and (ii) which network
characteristics and prior engagement characteristics are associated with
participation and effectiveness. We conduct a large scale field experiment with an
invitation design in which we manipulate both financial incentives and the soft
eligibility requirement to participate in the campaign. The latter provides a strong
and valid instrument to separate participation from outcomes effects. Since likes,
comments, and retweets are count variables, and since potential endorsers can
self-select to participate in multiple campaigns we ran, we use a Poisson
lognormal model with sample selection and correlated random effects to analyze
variations in participation and effectiveness. There are three main findings. (1)
Financial incentive did not affect either participation or effectiveness. (2) Potential
endorsers more likely to participate are often less effective. (3) Which
characteristics are associated with effectiveness depends on whether success is
measured in likes, comments, or retweets. Our findings provide new insights on
how marketers can improve social advertising campaigns by better targeting and
incenting potential endorsers.
Chair: Richard Sowers, Professor, University of Illinois at UrbanaChampaign, Urbana, IL, 61801, United States of America,
r-sowers@illinois.edu
1 - Model Uncertainty in Commodity Markets
Sebastian Jaimungal, University of Toronto,
Department of Statistical Sciences, Toronto, Canada,
sebastian.jaimungal@utoronto.ca, ¡lvaro Cartea, Zhen Qin
This paper studies the effect that model ambiguity in commodities have on the
value of derivative contracts. The base model consists of three drivers: a meanreverting diffusion, a mean-reverting jump, and a stochastic volatility component.
We allow the agent to consider a wide class of alternate models, and penalize the
differing components of the model individually. We demonstrate how agents alter
their behavior in the presence of ambiguity and how derivatives and spreads are
affected by it.
2 - Index Tracking with Futures
Brian Ward, Columbia University, New York, NY,
bmw2150@columbia.edu, Tim Leung
4 - A Visual Monitoring Technique Based on Importance Score and
Twitter Feeds
Zhenhuan Sui, Graduate Research Assistant, The Ohio State
University, 2501 Gardenia Drive, Columbus, Oh, 43235, United
States of America, sui.19@osu.edu, David Milam, Theodore Allen
Exchange Traded Funds (ETF) market continues to grow. Many ETFs are designed
to track an index, and their leveraged benchmarks (2x, 3x, etc.) Since many of
these indices are not directly traded, we consider tracking them using futures of
various maturities. We do so both statically and dynamically, and backtest our
strategies with empirical data.
We propose a visual monitoring technique based on topic models, a point system,
and Twitter feeds for monitoring. The method generates a chart showing the
important and trending topics that are discussed over a given time period which is
illustrate the methodology using cyber-security cases.
3 - Investment in Commodities ETFs and Management of Contango
Andrew Papanicolaou, NYU Polytechnic, 6 Metrotech Center,
Brooklyn, Ny, 11201, United States of America,
apapanic.brown@gmail.com
■ TC05
The last two decades have seen growing investment in commodities markets.
Commodities ETFs are popular but not a passive investment, as they will post
losses in contango markets. The focus of this talk will be storable commodities,
where uncertainty in the convenience yield reduces the Sharpe ratios. Losses are
seen as an information premium, which is quantified through a Merton-type
control problem.
Social Media Impact
Invited Session
Chair: Les Servi, The MITRE Corporation, 202 Burlington Road,
Bedford, MA, United States of America, lservi@mitre.org
1 - Big Data Means Big PR: A Review of News Coverage of Big Data
in the Popular Press
Amir Gandomi, Assistant Professor, Ryerson University,
350 Victoria Street, Toronto, ON, M5B 2K3, Canada,
agandomi@ryerson.ca, Murtaza Haider
319
TC07
4 - Speculative Oil
Anna Kruglova, Research Affiliate, MIT Center for Finance
and Policy, 30 Memorial Drive, Cambridge, MA, 02142,
United States of America, Kruglova@mit.edu, Andrei Kirilenko
2 - Free Riding in Team Projects: The Role of the Leadership Style
Morvarid Rahmani, Assistant Professor, Georgia Tech,
morvarid.rahmani@scheller.gatech.edu, Uday Karmarkar,
Guillaume Roels
The long-standing framework predicated on a premise that producers are the
main drivers of energy prices, the so-called “hedging pressure” theory, has been
shown to be less consistent with the empirical regularities present in the oil
prices. We hypothesize that with the influx of financial investors, the last needed
barrel is traded not between a hedger and a speculator, but between two
speculators: a commodity trader and/or a bank. We use granular information on
shipments of seaborne crude oil into the US during 2008-2012 to examine the
industry structure and determine who holds the crude oil that supports the
determination of prices in financial markets.
In order to remain innovative in today’s global market, firms are increasingly
organizing work around teams. In this paper, we investigate the role of the
leadership style (autocratic or democratic) on free-riding in teams and
characterize which leadership style is the most efficient depending on the project
characteristics.
3 - Different Approaches to Crowdfunding: Kickstarter vs. Indiegogo
Simone Marinesi, Wharton, 562 Jon M. Huntsman Hall, 3730
Walnut St, Philadelphia, PA, 19104, United States of America,
marinesi@wharton.upenn.edu, Karan Girotra
We compare the different modes of interaction between backers and creators
offered in the two most famous crowdfunding platforms, and provide
prescriptions on their implementation, taking the view of project creators.
■ TC07
4 - Are Good Idea Generators also Good at Evaluating Ideas
Otso Massala, INSEAD, 1 Ayer Rajah Avenue, Singapore,
Singapore, Otso.MASSALA@insead.edu
Systemic Risk: Methods and Models
Invited Session
Using data collected from a series of innovation tournaments we relate different
business opportunity generation skills with evaluation skills. We find that prolific
generators are worse evaluators while generators that produce high quality ideas
tend to also be good at evaluating. We provide implications for design of
innovation tournaments and innovative organizations.
Chair: David Brown, Duke University Fuqua School of Business,
1 Towerview Rd, Durham, NC, United States of America,
dbbrown@duke.edu
1 - Time-varying Systemic Risk: Evidence from a Dynamic Copula
Model of CDS Spreads
Dong Hwan Oh, Economist, Federal Reserve Board, 20th Street
and Constitution Avenue N.W., Washington, DC, 20551, United
States of America, donghwan.oh@frb.gov, Andrew Patton
■ TC09
Crowd Innovation
This paper proposes a new class of copula-based dynamic models for highdimensional conditional distributions, facilitating the estimation of a wide variety
of measures of systemic risk. We use the proposed new models to study a
collection of daily credit default swap spreads on 100 U.S. firms. We find that
while the probability of distress for individual firms has reduced since the
financial crisis of 2008-09, the joint probability of distress is substantially higher
now than in the pre-crisis.
Sponsored Session
Chair: Mohamed Mostagir, Assistant Professor, University of Michigan
Ross School of Business, 701 Tappan Ave, R5316, Ann Arbor, MI,
48109, United States of America, mosta@umich.edu
1 - Time-Based Crowdsourcing Contests
Ersin Korpeoglu, Carnegie Mellon University,
5000 Forbes Avenue, Pittsbugh, PA, United States of America,
ekorpeog@andrew.cmu.edu, Laurence Ales, Soo-Haeng Cho
2 - An Optimization View of Financial Systemic Risk Modeling
Nan Chen, Prof, Chinese University of Hong Kong, 709A William
Mong Engineering Building, Hong Kong, Hong Kong - PRC,
nchen@se.cuhk.edu.hk, Xin Liu, David D. Yao
We develop an optimization-based formulation on financial systemic risk. A
partition algorithm is constructed to solve the problem. The sensitivity analysis
helps us identify two multipliers to characterize the amplification effects caused
by liability networks and market liquidity. The effects of policy intervention are
also discussed in the paper.
We study a crowdsourcing contest wherein a seeker solicits a population of agents
to solve a problem. Each agent’s stochastic solution time depends on her effort
and expertise. We show that it is optimal for the seeker to screen and compensate
only the highest-expertise agents when their solution times are less uncertain, but
a larger group of agents when they are highly uncertain. An agent’s optimal
compensation is based on her solution time and whether the seeker can observe
agents’ efforts.
3 - Optimal Capital Requirements in Interbank Networks with Fire
Sales Externality
Jongsoo Hong, Duke University, 1 Towerview Rd, Durham,
NC, 27707, United States of America, jongsoo.hong@duke.edu,
David Brown
2 - Payoffs in Contests
Kevin Boudreau,Harvard University and London Business
School, Harvard Business School, Cambridge MA,
United States of America, kboudreau@hbs.edu, Karim Lakhani,
Nichale Menietti
We consider an interbank network with fire sales externality and study the
problem of optimally trading off between capital reserves and systemic risk. We
find that the optimal capital requirements is a solution to a stochastic linear
programming without fire sales externality and a stochastic mixed integer
programming with fire sales externality. We offer an iterative algorithm that
converges to the optimal. We demonstrate the methods on an example using data
from a central bank.
Many tournament outcomes possess signaling value. In this article, the results of
a field experiment on signaling incentives are presented. Using a structural model,
we obtain estimates of the value of nominal prizes, as well as extra value
associated with the contest. Signaling and cash values exhibit large interaction
effects. In all conditions, the perceived value of the prizes differs from the
nominal value. Competitors tend to undervalue small prizes and overvalue large
prizes.
■ TC08
3 - Achieving Efficiency in Dynamic Contribution Games
George Georgiadis, Assistant Professor, Northwestern University,
2001 Sheridan Rd, Evanston, IL, 60208, United States of America,
g-georgiadis@kellogg.northwestern.edu, Jaksa Cvitanic
Different Facets of Innovation: Product, Technology
and Business Models
We analyze a dynamic contribution game, in which a group of agents exert costly
effort over time to make progress on a project that generates a lump-sum payoff
once the cumulative effort reaches a pre-specified threshold. We characterize a
budget balanced mechanism which overcomes the free-rider problem, and at
every moment, induces each agent to exert the first-best effort level in a Markov
Perfect Equilibrium. Applications include early-stage entrepreneurial ventures
and joint R&D ventures.
Invited Session
Chair: Serguei Netessine, Professor, INSEAD, 1 Ayer Rajah Avenue,
Singapore, 138676, Singapore, Serguei.Netessine@insead.edu
1 - Identifying and Analyzing Styles in Design Patents
Tian Chan, INSEAD, TianHeong.CHAN@insead.edu,
Jurgen Mihm, Manuel Sosa
We introduce an approach to identify styles (categories of product designs similar
in form) among 400,000 US design patents. We combine state-of-the-art
clustering techniques with experimental validation to create, for the first time, a
dataset of styles. Building on this platform, we find that i) the level of turbulence
(unpredictability of changes) in styles follows a U-shaped pattern to the level of
turbulence in functionality, and ii) styles turbulence is increasing over time.
320
TC12
■ TC11
4 - Creativity under Fire: The Effects of Competition on
Creative Production
Daniel P. Gross, Post-doctoral Fellow, NBER/Harvard Business
School, Soldiers Field Road, Boston, MA, 02163,
United States of America, dgross@hbs.edu
Advances in Discrete Optimization
Sponsored Session
This paper studies the incentive effects of competition on individuals’ creative
production. Using a sample of commercial logo design competitions, and a novel,
content-based measure of originality, I find that competition has an inverted-U
effect on creativity. The results reconcile conflicting evidence from an extensive
literature on the effects of competition on innovation, with implications for R&D
policy, competition policy, and organizations in creative or research industries.
Chair: Gustavo Angulo, Assistant Professor, Pontificia Universidad
Catolica de Chile, Vicuña Mackenna 4860, Macul, Santiago, 7820436,
Chile, gangulo@ing.puc.cl
1 - Optimization over Structured Subsets of Positive Semidefinite
Matrices via Column Generation
Sanjeeb Dash, IBM Research, Yorktown Heights, NY,
United States of America, sanjeebd@us.ibm.com,
Amir Ali Ahmadi, Georgina Hall
■ TC10
e-Media and Health Care Practices
We describe LP and SOCP algorithms that optimize over some structured subsets
of the cone of positive semidefinite matrices (PSD cone) in an iterative fashion via
column generation, starting with an initial linear approximation of the PSD cone
given by Ahmadi and Majumdar (2014). We apply our techniques to sum-ofsquares programming for nonconvex polynomial optimization problems, and to a
copositive programming relaxation of the stable set problem.
Sponsor: E-Business
Sponsored Session
Chair: Harpreet Singh, University of Texas at Dallas, 800 West Campbell
Road, Richardson, United States of America, Harpreet@utdallas.edu
1 - Does the Adoption of EMR Systems Inflate
Medicare Reimbursements?
Kartik Ganju, Temple University, Fox School of Business,
Philadelphia, PA, United States of America,
tuc67632@temple.edu, Hilal Atasoy, Paul Pavlou
2 - Cutting Planes from Extended LP Formulations
Merve Bodur, UW-Madison, 1513 University Avenue,
mbodur@wisc.edu, Sanjeeb Dash, Oktay Gunluk
For mixed-integer sets, we study extended formulations of their LP relaxations.
We show that applying split cuts to such extended formulations can be more
effective than applying split cuts to the original formulation. For any 0-1 mixedinteger set with n integer and k continuous variables, we construct an extended
formulation with 2n+k-1 variables whose split closure is integral. We extend this
to general mixed-integer sets and construct the best extended formulation with
respect to split cuts.
We study if the adoption of the CPOE system is associated with an increase in the
complexity of the case mix that hospitals report (upcoding). We use the staggered
roll-out of the Recovery Audit Program as a natural experiment to assess the
impact of the adoption of the CPOE systems on the case mix that a hospital
reports. We find that the adoption of CPOE systems is associated with an increase
in the reported case mix of hospitals but the Audit program has had an effect on
reducing this.
3 - Robust (MONOTONE) Submodular Function Maxmization
Rajan Udwani, ORC, MIT, 70 Pacific Street, 324C, Cambridge,
MA, 02139, United States of America, rudwani@mit.edu,
James Orlin, Andreas Schulz
2 - Profit Complementarities in Adopting Electronic Medical
Records by U.S. Hospitals
Jianjing Lin, University of Arizona, 1130 E Helen St, Tucson, AZ,
85719, United States of America, jianjingl@email.arizona.edu
Consider two common instances of monotone submodular function maximization
with cardinality constraint, feature selection (machine learning) and sensor
placement. In both, it is often the case that out of the chosen set of features
(sensors), some may be corrupt (may fail). Thus, we would like our chosen set to
be robust to removal of some elements. We consider a previously known
formulation of this problem and give the first constant factor approximation
algorithms.
This paper tries to examine by how much hospitals’ profit can be increased if they
choose the locally market-leading vendor. Using a nationwide sample of U.S.
hospitals from 2006 to 2010, I construct a dynamic oligopoly model and recover
the model primitives with a classic approach in Economics. I find hospitals gain
significant benefits from choosing the local leader and if hospitals were
incentivized to choose such a technology, it would help improve the market
coordination substantially.
4 - On a Semicontinuous Relaxation of Fixed-charge Network
Flow Problems
Gustavo Angulo, Assistant Professor, Pontificia Universidad
Católica de Chile, Vicuña Mackenna 4860, Macul, Santiago,
7820436, Chile, gangulo@ing.puc.cl
3 - Get in Shape with Online Friends: Obesity and Social Networks
Behnaz Bojd, University of Washington, Seattle WA,
United States of America, behnaz@uw.edu, Yong Tan, Lu Yan
Obesity is one of the most prevalent health problems in the world. Individuals
suffering from this condition may use online social networks to get medical
information or emotional support. In this study, we examine the effects of online
social networking on individual weight-loss behavior, using a unique data set
from a platform where people can track their weight, seek food and fitness
information, share experiences, find buddies, and participate in different
challenges.
Usual formulations of fixed-charge network flow problems make use of binary
variables to indicate whether an arc is open or not, and to impose lower and
upper bounds on the flow whenever an arc is used. We propose a relaxation
where both binary and flow variables are treated as unbounded semicontinuous
variables. We derive a complete linear description of the convex hull of this
relaxation and show the tractability of the associated separation problem.
4 - Digital Word-of-mouth and Consumer Demand for Credence
Services: Evidence from a Natural Experiment
Aishwarya Deep Shukla, PhD Student, University of Maryland,
College Park, 3330C R H Smith School of Business, College Park,
20742, United States of America, adshukla@rhsmith.umd.edu,
Ritu Agarwal, Gordon Gao
■ TC12
Optimization Integer Programming I
Contributed Session
This paper examines the impact of online word-of-mouth on consumer demand
for credence services. We utilize a natural experiment setting from one of the
largest physician appointment booking platforms in India, when the website
made the doctor “recommendations” visible to users. In addition, we explore the
long term effect of this visibility of recommendations in the context of matching
the severity of the patient to the skill of the doctor.
Chair: John Chinneck, Professor, Carleton University, Systems and
Computer Engineering, 1125 Colonel By Drive, Ottawa, ON, K1S5B6,
Canada, chinneck@sce.carleton.ca
1 - School Districting Problem (SDP) Framed as a Spatial
Optimization Model/mixed Integer Program
Shawn Helm, Senior Manager, Portland Public Schools, 501
North Dixon Street, Portland, OR, 97227-1807, United States of
America, shelm@pps.net, Will Kearney, Sahan Dissanayake
SDPs assign neighborhoods to schools given physical, demographic, and policy
constraints. We control capacity; compactness; contiguity; amount of
demographic change. Users specify class size; grade band; neighborhood capture
assumptions; room uses; which current assignments are retained. The integrated
framework uses real GIS data from Portland Public Schools and available solvers
to identify best school-neighborhood assignments; solves model to display results
and inform boundary decisions.
321
TC13
2 - Facility Location Problem with Appointment Scheduling
in Healthcare
Mengnan Chen, University of Central Florida, Department of
IEMS, Orlando, FL, 32816, United States of America,
cmn891127@knights.ucf.edu, Qipeng Zheng
3 - Large Submatrix Detection in Gaussian Random Matrices
Quan Li, Massachusetts Institute of Technology, 550 Memorial
Drive Apt. 12B4, Cambridge, MA, 02139, United States of
America, quanli@mit.edu, David Gamarnik
Iterative Search Algorithm is widely used to find large average submatrices of a
real-valued matrix in the exploratory analysis. It alternately updates rows and
columns until no further improvement is obtained. We present first theoretical
analysis of its performance in Gaussian random matrices. We show that the
algorithm terminates within finite iterations w.h.p.. This result implies that it
converges to a local maximum submatrix w.h.p., leading to a constant factor gap
from the global maximum.
Facility location problem with appointment scheduling in healthcare is used to
schedule elective surgeries for physician, as well as to provide multiple choices for
patients. Facility location problem aims to improve the match between healthcare
resources (physician, clinic location) and patient needs (preferences and types of
diseases). By solving this problem, we can meet the minimum of loss for the
hospital (the total travel time) and satisfied the each patient preference as much
as possible.
4 - Locality in Optimization
Patrick Rbeschini, Yale University, New Haven, CT, United States
of America, Patrick.rebeschini@yale.edu, Sekhar Tatikonda
3 - Exploiting Variability with Machine Learning Based Restart
Strategies for MIP Solvers
Lars Beckmann, University of Paderborn, Warburger Strafle 100,
Paderborn, De, 33100, Germany, lars.beckmann@gmail.com
How does the solution of a constrained network optimization problem behave
under perturbations of the constraints with respect to the topology of the
network? Typically, sensitivity results concern the objective function evaluated at
the optimal point, not the optimal point itself. We develop a general theory for
the local sensitivity of optimal points on networks. In the context of the network
flow problem, this theory yields that local perturbations on the constraints have
an impact on the components of the optimal point that decreases exponentially
with the graph-theoretical distance. These results suggest a notion of decay of
correlation for (non-random) optimization procedures. This notion can be used to
develop local algorithms that can substantially reduce the computational
complexity of canonical optimization procedures.
Performance variability, a.k.a. the heavy-tail phenomenon, can greatly affect the
solution time of combinatorial problems, especially mixed-integer programming
(MIP) models. We present a machine learning based restart strategy that exploits
variability for quickly solving MIP problems. We show that this approach is
effective on a large dataset of MIP models. Our approach can be integrated in any
MIP solver and can potentially be generalized for solving other combinatorial
problems as well.
4 - Faster Infeasibility Analysis for Mixed Integer Linear Programming
John Chinneck, Professor, Carleton University, Systems and
Computer Engineering, 1125 Colonel By Drive, Ottawa, ON,
K1S5B6, Canada, chinneck@sce.carleton.ca, Andrew Scherr
■ TC14
Finding an Irreducible Infeasible Subset of Constraints in an infeasible mixed
integer linear program is extremely time consuming due to the need to solve
many different MILPs. We present a number of new algorithms that greatly speed
the process of analyzing infeasibility by reducing the number and size of MILP
subproblems solved.
Stochastic Financial Optimization
Sponsored Session
5 - Algorithms for Determining Internal Credit Ratings
Srinivas Bollapragada, Chief Engineer, General Electric,
1 Research Circle, Niskayuna, NY, 12309, United States of
America, bollapragada@ge.com, Xing Wang
Chair: Miguel Lejeune, Associate Professor, The George Washington
University, 2201 G Street, NW, Funger Hall, Suite 415, Washington, DC,
20052, United States of America, mlejeune@gwu.edu
1 - Scenario Decomposition for a Class of Risk-averse
Stochastic Programs
Pavlo Krokhmal, University of Iowa, 2136 Seamans Center,
Iowa City, IA, United States of America,
krokhmal@engineering.uiowa.edu
Financial Institutions assign credit ratings to customers for managing risks.
Underwriters use these ratings to determine risk-based pricing. We developed an
efficient algorithm to assign credit ratings to customers. Our algorithm partitions
the probability of defaults associated with customers into a specified number of
risk classes to achieve a target mean probability of default for each class. GE
Capital Services uses our algorithm for its risk management needs.
We consider nonlinear convex stochastic optimization problems where objective
or constraints involve downside coherent or convex risk measures of special form.
A scenario decomposition algorithm that exploits the constraint structure induced
by the corresponding risk measures is proposed. Numerical experiments on
portfolio optimization problems illustrate the computational effectiveness of the
developed procedure.
■ TC13
Stochastic Combinatorial Optimization
2 - Risk-budgeting Multi-portfolio Optimization with Portfolio and
Marginal Risk Constraints
Ran Ji, PhD Candidate, The George Washington University, 2201
G St, NW, Funger 415H, Washington, DC, 20052, United States of
America, jiran@gwmail.gwu.edu, Miguel Lejeune
Sponsored Session
Chair: Sean Skwerer, Yale, 300 George Street, Suite 523, New Haven,
CT, United States of America, sean.skwerer@yale.edu
1 - Finite Horizon Markov Decision Problems and a Central Limit
Theorem for Total Reward
Alessandro Arlotto, Duke University, Durham, NC,
United States of America, aa249@duke.edu, J. Michael Steele
We propose several stochastic risk budgeting multi-portfolio optimization models
with portfolio and marginal risk constraints. The models permit the simultaneous
optimization of multiple sub-portfolios with a downside risk measure assigned to
each asset and sub-portfolio. Each model includes a joint chance constraint with
random technology matrix. We expand a combinatorial modeling framework to
represent the feasible set of the chance constraint as a set of mixed-integer linear
inequalities.
We prove a CLT for a class of additive processes that arise naturally in the theory
of finite horizon Markov decision problems (MDPs). The theorem generalizes a
result of Dobrushin for temporally non-homogeneous Markov chains. The
innovation is that here the summands are permitted to depend on the current
state and a bounded number of future states of the chain. We show through
examples that this flexibility gives a direct path to asymptotic normality of the
total reward of finite horizon MDPs.
3 - Factor-based Robust Indexing
Roy Kwon, Associate Professor, University of Toronto,
5 King’s College Road, Toronto, Canada, rkwon@mie.utoronto.ca,
Dexter Wu
We present an approach for tracking a benchmark index using robust factor
models. Robust versions of the Fama-French 3 and 5 factor models are developed
to construct uncertainty sets for a robust conic integer program. Constraints limit
risk, tracking error, and number of tickets. A Lagrangian-based strategy is
developed and computational results in tracking the SP 100 and SP 500 show that
robust indexing can offer enhanced indexation in turbulent market conditions.
2 - Simplifying Ensembles of Trees
Sean Skwerer, Yale, 300 George Street, Suite 523,
New Haven, CT, United States of America,
sean.skwerer@yale.edu
Recursive partition functions (RPFs) are used in a wide variety of statistical
methods including classification and regression trees, regression splines, random
forests and boosting. The focus of this research is to find a framework for
aggregating an ensemble into a single tree or a small number of trees which have
comparable predictive strength to the entire ensemble. I will discuss advances
made in aggregating ensembles of recursive partition functions.
322
■ TC15
TC17
incorporates uncertainty in demand and presents a real-time stochastic
production plan and scheduling framework. SDVI will be used to obtain the
equilibrium solution.
2 - A Mixed Cooperative Dual to the Nash Equilibrium
Bill Corley, Professor, The University of Texas at Arlington,
P.O. Box 19017, Arlington, TX, 76019, United States of America,
corley@uta.edu
Optimization Models in Radiotherapy Treatment
Planning
Sponsored Session
A mixed dual to the Nash equilibrium is defined for n-person games in strategic
form. This dual extends the Berge equilibrium from pure to mixed strategies so
that mutual cooperation is achieved for the expected payoffs. Conditions are
established for the existence of a dual equilibrium. However, it is shown that for
each n>2 there exists a game for which no dual equilibrium exists. This fact may
be interpreted as there are mathematical as well as sociological obstacles to
mutual cooperation.
Chair: Victor Wu, PhD Student, University of Michigan,
vwwu@umich.edu
1 - Tractable Approaches to Multiple-needle Radiofrequency Ablation
Shefali Kulkarni-thaker, Graduate Student, University of Toronto,
5 Kings College Road, Medical Operations Research Lab (RS304),
Toronto, ON, (416) 978, Canada, shefali@mie.utoronto.ca,
Dionne Aleman, Aaron Fenster
3 - Nash’s Continuous Transformation and a Smooth Homotopy
Method for Computing Nash Equilibrium
Yabin Sun, PhD, City University of Hong Kong, R5218, Academic
Building 2, Tat Chee Avenue, Kowloon, Hong Kong, Hong Kong PRC, yabinsun-c@my.cityu.edu.hk, Chuangyin Dang, Yin Chen
In radiofrequency ablation (RFA), needles are used to apply extreme heat to
tumors, eradicating cancerous cells. To optimize multiple-needle RFA treatments,
we first obtain needle trajectories and positions using minimum volume covering
sphere and ellipse formulations. Then, we optimize the heat delivery duration for
each needle using tractable approximations to several thermal damage models.
We discuss resulting clinical treatment quality for four 3D patient models.
A different procedure often results in the different selection of Nash equilibrium.
To prove the existence of Nash equilibrium, Nash defined a continuous
transformation. This paper applies Nash’s continuous transformation to develop a
smooth homotopy method by introducing just one extra variable. Starting from
any given totally mixed strategy profile, the method numerically follows a smooth
path that ends at a Nash equilibrium. Extensive numerical results show that the
method is very efficient.
2 - Robotic Path Finding Techniques in Stereotactic Radiosurgery
Treatment Optimization
Marlee Vandewouw, University of Toronto, 5 King’s College
Road, Toronto, Canada, marleev@mie.utoronto.ca,
Kimia Ghobadi, Dionne Aleman, David Jaffray
4 - When to Release Feedback in a Dynamic Tournament
Ruoyu Wang, PhD Candidate, Fuqua School of Business,
Duke University, 100 Fuqua Drive, Durham, NC, 27708,
United States of America, rw120@duke.edu, Brendan Daley
We investigate applying robotic path finding techniques to develop treatment
plans for Gamma Knife Perfexion where the radiation is delivered continuously.
We explore the use of simultaneous localization and mapping, combined with
heuristic exploration techniques, to generate a path. A mixed integer model is
then used to find the beam times for this selected path. We discuss the advantages
and challenges of this method in comparison to the conventional forward and
inverse step-and-shoot plans.
We study dynamic tournaments in which time is modeled explicitly, as opposed to
with the abstract notion of periods. By doing so, we characterize the effects of the
ex-ante-designated timing of an interim progress report. Whether a policy of
reporting increases total expected effort does not depend on the release time. We
find that total expected effort is single-peaked/single-troughed in the report’s
release time, with the peak/tough located at a time more than halfway through
the tournament.
3 - Adaptive and Robust Radiation Therapy in the Presence of Drift
Philip Allen Mar, Dept. of MIE, University of Toronto,
5 King’s College Road, Toronto, ON, M5S 3G8, Canada,
philip.mar@mail.utoronto.ca, Timothy Chan
5 - Endgame Solving in Large Imperfect-information Games
Sam Ganzfried, Carnegie Mellon University, Computer Science
Department, 5000 Forbes Avenue, Pittsburgh, PA, 15213, United
States of America, sam.ganzfried@gmail.com, Tuomas Sandholm
We present our computational study of an adaptive and robust optimization
radiation therapy (ARRT) method. Previously, it was shown that this ARRT
method provides asymptotically optimal treatment plans for convergent
sequences of tumor motion distributions. In this work, we generate simulated
sequences of tumor motion distributions that exhibit baseline, amplitude and
breathing phase drift, and show the effectiveness of the ARRT method applied to
these sequences.
Sequential games of perfect information can be solved in linear time by a
straightforward backward induction procedure; however, this procedure does not
work in games with imperfect information since different endgames can contain
nodes that belong to the same information set and cannot be treated
independently. We present an efficient algorithm for performing endgame solving
in large imperfect-information games and demonstrate its success experimentally
in two-player no-limit Texas hold ‘em.
4 - Vmat Radiation Therapy: Modeling Treatment Delivery Time
Versus Plan Quality
David Craft, Massachusetts General Hospital, 30 Fruit St, Boston,
MA, 02114, United States of America, dcraft@alum.mit.edu,
Marleen Balvert
■ TC17
Volumetric modulated arc therapy is a radiation method where the gantry
delivers dose continuously as it rotates around the patient. Metal leaves sweep
across the field to modulate the intensity fields. In commercial software, leaf
trajectories are solved by heuristics without any guarantee of an optimality gap.
VMAT is a large scale non-convex optimization problem with many local minima.
We offer a solution approach and explore the tradeoff between treatment quality
and delivery time.
Network Analysis I
Sponsored Session
Chair: Alexander Veremyev, University of Florida, 1350 N Poquito
Road, Shalimar, FL, United States of America, averemyev@ufl.edu
1 - Optimizing Network Recovery Time under Uncertainty
Juan Borrero, University of Pittsburgh, 3700 O’Hara Street,
Pittsburgh, PA, 15213, United States of America, jsb81@pitt.edu,
Pavlo Krokhmal, Oleg Prokopyev
■ TC16
Game Theory I
We consider a network under attack, where its nodes can recover either on their
own, by receiving support from neighboring nodes, or by receiving support from
outside the network. A decision maker has to determine how to invest his budget
on these options in order to minimize recovery time. We propose a novel
hierarchical and stochastic model to address the issue, derive closed form
equations for the optimal resource allocation, and study its behavior as the
number of nodes grows to infinity.
Contributed Session
Chair: Sam Ganzfried, Carnegie Mellon University, Computer Science
Department, 5000 Forbes Avenue, Pittsburgh, PA, 15213, United States
of America, sam.ganzfried@gmail.com
1 - A Stochastic Approach for Dynamic Urban Supply
Chain Management
Afrooz Ansaripour, Pennsylvania State University, 244 Leonhard
building, State College, PA, United States of America,
afrooz.ansaripour2000@gmail.com, Wenjing Song,
Terry Friesz, Yiou Wang, Zhaohu Fan
Lack of information sharing causes negative impacts such as traffic and pollution.
City logistics aims to optimize urban freight systems. This paper is an extension of
recent stochastic vehicle routing and scheduling frameworks. These frameworks
do not necessarily account for real-time variability in traffic. This paper
323
TC19
2 - s-plex and s-defective Numbers of a Graph
Vladimir Stozhkov, University of Florida, 2330 SW Williston Rd
Apt. 2826, Gainesville, FL, 32608, United States of America,
vstozhkov@ufl.edu, Eduardo Pasiliao, Vladimir Boginski
problem. We use this property to design a heuristic procedure to improve the
quality of the initial solution.
4 - Renewable Energy Prediction and Prescription in the
Internet-of-things (IoT)
Hans Schlenker, IBM, Hollerithstr 1, Munich, 81829, Germany,
hans.schlenker@de.ibm.com, Yianni Gamvros
The presentation is dedicated to two clique relaxation models: s-plex and sdefective clique. Theoretical properties of the specified objects are investigated.
Analytical and computational bounds for the related optimization problems are
provided. The extensions of the Motzkin-Straus formulation for s-plex and sdefective clique are derived. The outline of the general procedure for solving the
corresponding maximization problems is given.
The IoT connects all sorts of devices — from sensors to embedded devices to
smartphones to laptops to servers. IBM connected 1600 solar fields to its
Renewable Energy IoT. Sensor data is collected, combined in the cloud, and
further analyzed by analytics services to generate accurate local energy
production forecast. These predictions are then used by (prescriptive)
mathematical optimization in a network distribution model to balance under-runs
and over-production in all connected areas.
3 - Minimum Edge Blocker Dominating Set Problem
Foad Mahdavi Pajouh, Assistant Professor, University of
Massachusetts Boston, 100 Morrissey Boulevard, Boston, MA,
02125, United States of America, Foad.Mahdavi@umb.edu,
Eduardo Pasiliao, Jose Walteros, Vladimir Boginski
Dominating sets are widely used in social and communication networks analysis.
Given a weighted graph and r>0, we consider the problem of removing a
minimum number of edges so that the weight of any dominating set in the
remaining graph is at least r. Complexity results, polyhedral results, a linear 0-1
programming formulation, and an exact algorithm for solving this problem will be
presented.
■ TC20
4 - Minimum Risk Network Covering Location Problem
Konstantin Pavlikov, University of Florida, 1350 N. Poquito Road,
Shalimar, FL, 32579, United States of America,
kpavlikov@ufl.edu, Alexander Veremyev, Vladimir Boginski,
Eduardo Pasiliao
Chair: Zhen Liu, Options Clearing Corp (OCC), One North Wacker
Drive, Suite 500, Chicago, IL, 60606, United States of America,
zhenliu@alum.northwestern.edu
1 - An Optimization Procedure for a Delta Neutral Constrained Theta
with Maximum Gamma Portfolio
Arik Sadeh, Dean, HIT Holon Institute of Technology,
52 Golomb St. 5810201, Holon, Israel, sadeh@hit.ac.il
Financial Engineering and Optimization
Contributed Session
The network coverage problem under uncertainty is considered. In this problem,
components of the covering set and links connecting them to remaining nodes of
the network are subject to random failures. The emphasis is put on minimizing
the risk of losing coverage in presence of such failures. We formalize the model
and discuss its connection to the maximum expected covering location model.
A large gamma portfolio is attractive for investors in order to get benefits from
large increase or decrease in the value of the underlying asset. In large gamma
portfolio the theta is negative. In this study, a delta neutral portfolio with
maximum gamma and constrained theta, was developed. An optimization model
was designed and solved for small time steps within a planning horizon. The
model was run for many simulation scenarios as well as real world data, followed
by statistical tests.
■ TC19
2 - Optimal Portfolio Liquidation and Dynamic
Mean-variance Criterion
Jiawen Gu, Postdoc, University of Copenhagen, Department of
Mathematical Science, University of Copenhagen, Copenhagen,
2100, Denmark, kaman.jwgu@gmail.com, Mogens Steffensen
Modeling and Optimization for Sustainable
Cloud Computing
Sponsored Session
We consider the portfolio liquidation problem under the dynamic mean-variance
criterion and derive time-consistent solutions in three important models. We get
explicit trading strategies in the basic model and when random pricing signals are
incorporated. When consider stochastic liquidity and volatility, we construct an
exact HJB equations under general assumptions for the parameters.
Chair: Yunpeng Pan, South Dakota State University, Mathematics &
Statistics, Box 2220, Brookings, SD, 57007, United States of America,
yunpeng.pan@gmail.com
1 - Remote Sensing Data Mining for Extracting Data Center
Site Characteristics
Yunpeng Pan, South Dakota State University, Mathematics &
Statistics, Box 2220, Brookings, SD, 57007, United States of
America, yunpeng.pan@gmail.com, Adam Buskirk
3 - Genetic Programming Optimization for a Sentiment Feedback
Strength Based Trading Strategy
Steve Yang, Assistant Professor, Stevens Institute of Technology,
1 Castle Point on Hudson, Hoboken, NJ, 07030,
United States of America, steve.yang@stevens.edu
Data centers are powerhouses of cloud. Companies rush to build out their cloud
infrastructure to meet fast growing demand. The environmental impact such as
carbon footprint falls into the category of public good, and therefore, calls for
appropriate public policy decisions, which in turn require good information. Our
current work intends to achieve this by mining the Landsat remote sensing data
to extract characteristics of data centers in operation and under construction at a
global scale.
Based on the evidence that tweets are faster than news in revealing new market
information, whereas news is regarded broadly a more reliable source of
information than tweets, we develop a trading strategy based on the sentiment
feedback strength between the news and tweets using generic programming
optimization method. Result shows that this strategy generates over 14.7%
Sterling ratio compared with 10.4% and 13.6% from the technical indicatorbased and the buy-and-hold strategy respectively.
2 - A Dynamic Workflow Framework for Server Provisioning
Wei Lin, Software Engineering Researcher, IBM,
8 Dongbeiwang Western Road, Haidian Dist, Beijing, China,
linweilw@cn.ibm.com, Brian Peterson, Qinhua Wang,
Zongying Zhang, Christopher Young, Sai Zeng
4 - Algorithmic Options Trading by Integer Programming
Vadim Timkovski, Keiser University, Port St. Lucie, FL,
United States of America, vtimkovski@keiseruniversity.edu
Algorithmic options trading has only begun its evolution. This work presents an
integer programming system that simulates the activities of an experienced option
trader on the construction and adjustment of option portfolios. The system adopts
algorithms based on a recent discovery of an algebraic classification of option
trading strategies, without which this kind of automation would not be possible
and which has not been considered before as attainable.
Cloud service providers support server provisioning to large number of enterprise
customers, who have different functional, security and compliance requirements.
We propose a framework which composes dynamic workflow at runtime to cater
individualized provisioning procedures. In this framework, an onboarding module
configures process steps and dependencies for each customer, and a composition
module dynamically composes execution workflow based on dependency
validation and sequence calculating.
5 - Linear Programming Approach to American Option Pricing
Zhen Liu, Options Clearing Corp (OCC), One North Wacker
Drive, Suite 500, Chicago, IL, 60606, United States of America,
zhenliu@alum.northwestern.edu
3 - Minimizing Costs in Distributed Cloud Resource Provisioning
Julio Goez, Postdoctoral Fellow, Ecole Polytechnique Montreal
and GERAD, 2900 Boulevard Edouard-Montpetit, Montréal, QC,
H3T 1J4, Canada, jgoez1@gmail.com, Juan F. Pérez
We solve the variational inequality (VI) from American option pricing problem by
linear programming (LP) approach. We approximate its solution by a combination
of Chebyshev basis functions. The objective is to minimize the absolute error of
the solution and the max operator in VI is converted into linear constraints of LP.
We discuss its convergence, and compare our results with Longstaff-Schwartz
least-square approach and numerical partial differential equation (PDE) approach.
We consider the problem of minimizing the cost of provisioning resources at
different cloud locations, constrained to satisfying a required service-level
objective. We present a mixed integer non-linear optimization model for this
problem and show an equivalent mixed integer second order cone formulation.
We also show that a simple round-up provides an initial feasible solution for the
324
■ TC21
TC23
substantial empirical evidence for slowdown, particularly when a patient’s delay
exceeds a certain threshold. For such threshold slowdown situations, we design
and analyze a many-server system that leads to a two-dimensional Markov
process. Analysis of this system leads to insights into the potentially detrimental
effects of slowdown.
Innovations in Healthcare Operations
Sponsored Session
3 - Maximizing throughput in Non-collaborative Networks of Queues
Tugce Isik, Georgia Institute of Technology, 755 Ferst Drive NW,
Atlanta, GA, 30332-0205, United States of America,
tugceisik@gatech.edu, Hayriye Ayhan, Sigrun Andradottir
Chair: Mili Mehrotra, University of Minnesota, 321 19th ave south,
minneapolis, United States of America, milim@umn.edu
1 - Incentizing Less-Than-Fully-Qualified Providers for Early
Diagnosis of Tuberculosis in India
Sarang Deo, Assistant Professor, Indian School of Business
Hyderabad, ISB Hyderabad, Gachibowli, Hyderabad, TS, 500032,
India, sarang_deo@isb.edu, Milind Sohoni, Neha Jha
We study queueing networks with flexible non-collaborative servers. We
introduce a processor sharing (PS) scheme that yields maximal throughput when
buffers are infinite. For systems where the servers cannot work together at a
station, we develop non-collaborative round-robin policies that approximate PS
as the rotation of the servers becomes more frequent. We evaluate the
performance of these policies in queueing networks with tandem, merge, and
split topologies for different buffer sizes.
A major driver of TB epidemic in India is delay in diagnosis by less-than-fullyqualified providers (LTFQs), who are typically the first point of contact for
patients. This work is motivated by pilots funded by international donors to
provide monetary incentives to LTFQs to induce earlier diagnosis. We develop a
game-theoretic model to design an incentive contract that should be offered to
LTFQs and calibrate it using realistic parameter estimates obtained from primary
and secondary data.
4 - Optimal Assignment of Authentication Servers to Different
Customer Classes
Daniel Silva, Georgia Tech, 755 Ferst Drive, Atlanta, GA, United
States of America, dfsi3@gatech.edu, Hayriye Ayhan, Bo Zhang
Consider a system where user requests for authentication arrive from several
classes of customers, following independent Poisson processes. Each arrival has a
class-dependent probability of being an impostor. The system has several
authentication methods; each one has a known service time distribution, and a
Type I and II error probability. A controller assigns a method to each user request.
We model the system as a queueing network and find the structure of a costoptimal routing policy.
2 - Optimizing Spatiotemporal Antiviral Release Schedules in a
Pandemic Influenza
Bismark Singh, Assistant Professor, University of Texas at Austin,
Austin, TX, United States of America, ned@austin.utexas.edu,
Nedialko Dimitrov
To help the state of Texas plan influenza pandemic interventions, we build a
stochastic MIP to compute time-based antiviral releases. We derive scenarios for
the stochastic program from an epidemic simulator that accounts for the large
amount of uncertainty in disease progression. We study the hardness of this
problem, and present models and methods to solve it, even though a direct-solve
is intractable because of the large number of scenarios.
■ TC23
Stochastic Modeling and Control of Production
Systems
3 - Online Scheduling of Operating Rooms
Chaitanya Bandi, Kellogg School of Management,
Northwestern University, Evanston, IL, United States of America,
c-bandi@kellogg.northwestern.edu, Diwakar Gupta
Invited Session
We consider the online operating room scheduling problem where we do not
know the sequence of requests and associated surgery lengths beforehand. Given
the uncertainty and the objective of feasible schedules, we model the uncertainty
using a Robust Optimization (RO) approach, and utilize a RO framework to
develop an interval-classification scheduling algorithm optimized under the RO
framework. We obtain provable lower bounds on the performance and show
promising results based on real data.
Chair: Sanket Bhat, McGill University, 1001 Sherbrooke Street West,
Room 520, Montreal, QC, H3A 1G5, Canada, sanket.bhat@mcgill.ca
1 - Using an Artificial Neural Network Model and Approximate
Dynamic Programming for Stochastic Control
Han Wu, Student, University of Louisville, 2301 S 3rd St,
Louisville, KY, 40218, United States of America,
han.wu@louisville.edu, Gerald Evans, Kihwan Bae
4 - Is Technology Eating Nurses? – Staffing Decisions in
Nursing Homes
Feng Lu, Assistant Professor, Purdue University, 403 W State St,
lu428@purdue.edu, Huaxia Rui, Abraham Seidmann
Development of efficient control policies for dynamic production systems is
difficult. The uncertain demands and large set up times on machines can cause
significant problems. Consider an assembly line for dishwashers which require
multiple types of wire racks that must be fabricated and coated at different
machines. An Artificial Neural Network model is embedded within an
approximate dynamic programming algorithm to search for a better production
and inventory control policy.
We study the effect of IT-enabled automation on staffing decisions in healthcare
facilities using a unique nursing home IT data from 2006 to 2012. We also
develop a strategic staffing model that incorporates technology adoption.
2 - Resource Allocation Policies to Provide Differentiated Service
Levels to Customers
Ananth Krishnamurthy, Associate Professor, University of
Wisconsin-Madison, 1513 University Avenue,, ME 3258,
akrishn2@wisc.edu, Sanket Bhat
■ TC22
Analysis and Control of Queues
We analyze resource allocation decisions for component manufacturers who
supply components to several original equipment manufacturers (OEMs). OEMs
differ in their demand variability and service level expectations. We derive
policies that provide differentiated service to OEMs depending on their demand
variability. Under the dynamic programming framework, we investigate the value
of these policies to component manufacturers.
Sponsored Session
Chair: Hayriye Ayhan, Georgia Tech, Atlanta, GA, United States of
America, hayriye.ayhan@isye.gatech.edu
1 - Control of Multiserver Energy-aware Queueing Systems
Vincent Maccio, McMaster University, 1280 Main Street West,
Hamilton, Canada, macciov@mcmaster.ca, Douglas Down
3 - A Newsvendor Problem with Price-sensitive and Uncertain Supply
Z. Melis Teksan, University of Florida, ISE Dept. 303 Weil Hall,
P.O. Box 116595, Gainesville, FL, 32611, United States of
America, zmteksan@gmail.com, Meltem Tutar, Joseph Geunes
We study the problem of controlling a multiple server system, where servers may
be turned on or off. The cost function of interest is a combination of holding costs
and energy costs (and potentially switching costs). We provide several structural
results on the optimal policy - these structural results are enough to allow for the
derivation of the optimal policy for a wide range of systems. Finally, we discuss
how these policies compare with those extant in the literature.
We study a newsvendor problem in which the supply quantity depends on the
price offered by the newsvendor to suppliers. We analyze the optimal ordering
policy, which depends on the economics of overage and underage costs, as well as
the relationship between price and supply quantity. We characterize the optimal
supply-pricing policies for cases in which suppliers are also unreliable, i.e., supply
capacity is both price-dependent and random.
2 - The Snowball Effect of Customer Slowdown in Critical
Many-server Systems
Jori Selen, PhD Candidate, Eindhoven University of Technology,
De Zaale, Eindhoven, Netherlands, j.selen@tue.nl,
Johan Van Leeuwaarden, Vidyadhar Kulkarni, Ivo Adan
Customer slowdown describes the phenomenon that a customer’s service
requirement increases with experienced delay. In healthcare settings, there is
325
TC24
4 - Optimal Control of an Inventory System with Stochastic and
Independent Leadtimes
Mohsen Elhafsi, Professor, University of California, School of
Business Administration, 900 University Avenue, Riverside, CA,
92521, United States of America, mohsen.elhafsi@ucr.edu,
Saif Benjaafar, Rui Chen
social media can be used to identify adverse drug events with supervised learning
approaches. However, they require expert annotation and are not scalable for
large datasets. In this study, we develop a framework for pharmacovigilance in
social media using distant supervision. Our framework achieves competent
performance without annotation.
2 - The Effect of Rating System Design on Negativity Bias
Ying Liu, Arizona State University, 1201 S McClintock Dr,
Apt 221, Tempe, AZ, 85281, United States of America,
yingliu_is@asu.edu, Pei-yu Chen, Kevin Hong
We study a continuous review inventory system with stochastic independent
leadtimes. Because orders may not be delivered in the same sequence in which
they have been placed, characterizing the optimal policy is difficult and much of
the available literature assumes a fixed base-stock policy which we show is suboptimal and can perform poorly. Instead, the optimal policy is state-dependent
and specified in terms of an inventory-dependent threshold function
characterized by at most m parameters.
Does rating system design affect consumers’ negativity bias in reporting product
ratings? We examine the effect with both observational and experimental study.
Results suggest that consumers tend to reflect their experiences in the least
satisfied dimension in single-dimensional rating systems, whereas the overall
ratings in multi-dimensional systems tend to reflect consumers’ average
experience. The study suggests that multidimensional rating systems could
mitigate consumers’ negativity bias.
■ TC24
3 - Predict Campaign Quality: An Empirical Analysis of the Value of
Video in Crowdfunding Markets
Qiang Gao, University of Arizona, 3700 N 1st Ave. #1020, Tucson,
AZ, 85719, United States of America, qiangg@email.arizona.edu,
Mingfeng Lin
Search Across Disciplines: Artificial Intelligence and
Operations Research
Sponsored Session
Videos are prevalent in crowdfunding campaigns where there is usually little
verifiable information. Yet to date there is virtually no systematic study of its roles
in this new context. We investigate how video features predict campaign quality
using data from a leading rewards-based crowdfunding website by implementing
both explanatory and predictive models.
Chair: Nathan Sturtevant, University of Denver, 2280 S. Vine St.,
Denver, CO, 80210, United States of America, sturtevant@cs.du.edu
1 - The Cyclic Best-first Search Strategy for
Branch-and-bound Algorithms
Jason Sauppe, University of Illinois at Urbana-Champaign, 201
North Goodwin Avenue, Urbana, IL, 61801, United States of
America, sauppejj@gmail.com, Edward Sewell, Sheldon Jacobson,
David Morrison
4 - Content Monetization in Social Media: Estimation of Demand and
Supply for User Generated Content
Ruibin Geng, Zhejiang University, 388 Yuhangtang Road,
Hangzhou, ZJ, 310058, China, grace.bin1207@gmail.com,
Bin Zhang, Paulo Goes
Social networking is reaching a maturity stage with fewer new registrations but
more user churning. Our study investigates how a new market mechanism,
content monetization, reduces turnover rate by using data from the largest
Chinese social network Sina Weibo. It examines the factors that affect both the
demand and supply for user-generated content (UGC) in social media. Our results
confirm that this nascent mechanism effectively motivates the supply for UGC
and also improves its quality.
The Cyclic Best-First Search (CBFS) strategy is a generalization of best-first search
that splits unexplored subproblems over a collection of heaps, referred to as
contours. During the search process, CBFS repeatedly cycles through a list of nonempty contours, selecting one subproblem to explore from each during every
pass. Contours can be defined in various ways to influence the search process.
This talk will present some properties of CBFS and computational results for a
variety of problems.
2 - Adding Random Exploration to Search Algorithms
Rick Valenzano, Alberta Innovates Centre for Machine Learning,
2-21 Athabasca Hall, University of Alberta, Edmonton, AB,
T5K1X4, Canada, valenzan@ualberta.ca
■ TC26
In this work, we will use a simple technique called epsilon-greedy node selection
to demonstrate the value of enhancing search algorithms with random
exploration. Through empirical evaluation, this technique is shown to
substantially improve the performance of search-based automated planners. We
also formally analyze this technique to demonstrate that algorithms that employ
random exploration are more robust to heuristic error.
Gray Market, Sustainability, Competition, and
Diffusion
3 - Exploiting Large Admissible Heuristics in Search
Nathan Sturtevant, University of Denver, 2280 S. Vine St.,
Denver, CO, 80210, United States of America,
sturtevant@cs.du.edu
Chair: Samar Mukhopadhyay, Professor, Sungkyunkwan UniversityGSB, 25-2 Sungkyunkwan-ro, Jongno gu, Seoul, 110 745, Korea,
Republic of, samar@uwm.edu
1 - Countering Gray Market Threat using Marketing Effort
Samar Mukhopadhyay, Professor, Sungkyunkwan UniversityGSB, 25-2 Sungkyunkwan-ro, Jongno gu, Seoul, 110 745, Korea,
Republic of, samar@uwm.edu, Xuemei Su
Invited Session
Admissible heuristics, which do not overestimate the cost to the goal, are
particularly useful in shortest-path search problems, as they can guide the search.
They also can, if necessary, guarantee optimal solutions. This talk looks at recent
heuristics that are larger than the memory, and suggests ways of exploiting the
problem structure to reduce the memory overhead of storing the heuristic. Two
methods, bloom filters and value range compression are discussed, along with
their tradeoffs.
Gray markets are likely when there is a significant price difference of the same
product in different markets. This paper studies the role of an important variable,
marketing effort, in fighting gray market, in addition to price. We find that when
both marketing effort and prices are controlled, the manufacturer’s profit is
improved. Sometimes, it may even be better not to sell through the authorized
channel, but to manage the gray market by controlling the marketing effort levels
and prices.
■ TC25
2 - Designing Sustainable Products under Co-Production Technology
Yen-Ting Lin, University of San Diego, School of Business
Administration, 5998 Alcala Park, San Diego, CA, United States of
America, linyt@sandiego.edu, Haoying Sun, Shouqiang Wang
Online Crowds: Crowdfunding and Social Media
Sponsored Session
We consider a manufacturer who takes a natural resource to make two products
through co-production technology. Some consumers are green and additionally
value conservation of the natural resource. We show that increasing the portion
of green consumers may actually elevate resource consumption.
Chair: Qiang Gao, University of Arizona, 3700 N 1st Ave. #1020,
Tucson, AZ, 85719, United States of America,
qiangg@email.arizona.edu
1 - A Distant Supervision Approach for Social Media
Pharmacovigilance
Xiao Liu, University of Arizona, 1130 E. Helen St.,
Room 430, Tucson, AZ, 85721, United States of America,
xiaoliu@email.arizona.edu, Hsinchun Chen
Phamacovigilance refers to the science relating to the detection, assessment,
understanding, and prevention of adverse drug events. Prior studies showed
326
TC29
■ TC28
3 - Online Vs. Traditional Education: A Competitive Framework.
Vashkar Ghosh, University of Florida, Department of ISOM,
Gainesville, FL, 32611, United States of America,
vashkar.ghosh@warrington.ufl.edu, Gulver Karamemis,
Asoo Vakharia
New Frontiers in Market Design
Cluster: Auctions
Invited Session
Innovation and technological advancement are eliminating a lot of constraints
(eg. physical presence) bringing sweeping changes to higher education. We
examine how technology in higher education is likely to develop and what its
impacts will be on existing institutions. We examine a university’s incentive to
offer online programs in addition to the traditional program in a competitive
environment. We consider two different games: the simultaneous and the
sequential leader/follower location game.
Chair: Tunay Tunca, ttunca@rhsmith.umd.edu
1 - Integrating Market Makers, Limit Orders, and Continuous Trade in
Prediction Markets
Sebastien Lahaie, Microsoft Research, New York, NY, United
States of America, sebastien.lahaie@gmail.com, Hoda Heidari,
David Pennock, Jenn Wortman Vaughan
4 - The Diffusion of Product Generation of Auto Industry
Gary Chao, Kutztown University, P.O. Box 730, Kutztown, PA,
18031, United States of America, chao@kutztown.edu,
Maxwell Hsu
We provide an algorithm that combines market makers and limit orders in a
prediction market with continuous trade. We define the notion of an approximate
trading path, a path in security space along which orders execute at their limit
prices to within a fixed tolerance. We show that a trading path with efficient
endpoint exists under supermodularity, but not in general. We develop an
algorithm for the general case, and evaluate it using real combinatorial
predictions over election outcomes.
Instead of a whole new model, every a few years, automakers introduce a new
generation of their existing model to continue their success of old models or to
correct the mistakes in the old models. Based on the Bass diffusion theory, we
would like to study whether the different sales behavior among models and
generations in US market.
2 - Multi-dimensional Virtual Values and Second-degree
Price Discrimination
Nima Haghpanah, MIT, Boston, MA, United States of America,
nima.haghpanah@gmail.com, Jason Hartline
■ TC27
We consider a problem of selling a product with multiple quality levels and derive
conditions that imply only selling highest quality is optimal. With multidimensional preferences, virtual values from integration by parts on arbitrary
paths may not be incentive compatible. To resolve this issue, we impose
additional conditions that are satisfied only by a unique choice of paths, and
identify distributions that ensure the resulting virtual surplus is indeed point-wise
optimized by the mechanism.
Evolutionary Bilevel Optimization
Sponsored Session
Chair: Kalyanmoy Deb, Koenig Endowed Chair Professor, Michigan
State University, 428 S. Shaw Lane, 2120 EB, East Lansing, MI, 48864,
United States of America, kdeb@egr.msu.edu
1 - Bilevel Decision Making and Optimization
Pekka Malo, Assistant Professor, Aalto University School of
Business, Runeberginkatu 22-24, Helsinki, Finland,
pekka.malo@aalto.fi, Ankur Sinha, Kalyanmoy Deb,
Jyrki Wallenius, Pekka Korhonen
3 - Optimal Pricing for Two-sided Platforms with Externalities
Levi Devalve, Duke University, Durham, NC,
United States of America, levi.devalve@duke.edu, Sasa Pekec
We consider pricing strategies of two-sided platforms serving consumers and
marketers. We show that competing platforms can achieve optimal profit through
“subscription-only” pricing. We identify settings in which competition increases
both consumer prices and advertising volumes. We also derive the platform’s
optimal price menu under incomplete information about the consumer’s disutility
for advertising. We characterize when the optimal menu includes free use and no
ad options.
Bilevel decision making and optimization problems are commonly framed as
leader-follower problems, where the leader desires to optimize his own decision
while taking the decisions of the follower into account. In such cases, the Paretooptimal frontier of the leader is influenced by the decision structure of the
follower facing multiple objectives. In this paper, we analyze this effect by
modeling the lower level decision maker using value functions.
4 - The Role of a Market Maker in Networked Cournot Competition
Desmond Cai, California Institute of Technology, 1200 E
California Blvd, Pasadena, CA, 91125, wccai@caltech.edu,
Subhonmesh Bose, Adam Wierman
2 - Handling Uncertainties in Decision Variables for Bilevel
Optimization Problems
Kalyanmoy Deb, Koenig Endowed Chair Professor, Michigan
State University, 428 S. Shaw Lane, 2120 EB, East Lansing, MI,
48864, United States of America, kdeb@egr.msu.edu, Zhichao Lu
We study the role of a market maker (or market operator) in a transmission
constrained electricity market. We model the market as a one-shot networked
Cournot competition. We analyze the class of market maker objective functions
given by linear combinations of social welfare, residual social welfare, and
consumer surplus. We show that there exist cost functions for which the
maximum possible social welfare at equilibrium is not attained when the market
maker chooses to maximize social welfare.
Bilevel problems involve two optimization problems in hierarchy and are
challenging problems often found in practice. In this talk, we present
evolutionary optimization algorithms and results on test and practical bilevel
problems with uncertainties in decision variables for finding robust and reliable
solutions. Uncertainties are considered for both lower and upper level variables
and problems with and without constraints.
■ TC29
3 - Expected Frontiers: Incorporating Weather Uncertainty into an
Integrated Bilevel Optimization
Moriah Bostian, Assistant Professor, Lewis & Clark College,
Department of Economics, 0615 SW Palatine Hill Rd, Portland,
OR, 97219, United States of America, mbbostian@lclark.edu,
Gerald Whittaker, Bradley Barnhart, Rolf Fare, Shawna Grosskopf
Joint Session Analytics/HAS:
Analytics Innovations in Healthcare and Medicine
Sponsor: Analytics
Sponsored Session
Weather is a main driver of agricultural nutrient fate and transport in the
environment. We use bilevel optimization and a time-series bootstrap to evaluate
a water pollution policy subject to a distribution of weather outcomes. Our results
show that the deterministic Pareto frontier is sensitive to climate variation. Some
policy configurations that appear equally effective in a deterministic model setup
are strongly differentiated when weather uncertainty is included in the policy
evaluation.
Chair: Issac Shams, Postdoctoral Research Fellow, University of
Michigan, 1205 Beal Ave, Ann Arbor, United States of America,
issacsh@umich.edu
1 - Improving Societal Outcomes in the Organ Donation Value Chain
Priyank Arora, Georgia Institute of Technology, 800 W Peachtree
St. NW, Atlanta, GA, 30308, United States of America,
priyank.arora@scheller.gatech.edu, Ravi Subramanian
We examine a unique principal-agent problem in the cadaver organ donation
value chain (ODVC) where the principal in our case is a social planner that has an
overall quality-adjusted-life-year improvement objective. The agents include a
non-profit organ procurement organization with a volume-of-care objective and a
for-profit hospital (trauma center).
327
TC30
■ TC31
2 - Appointment Scheduling and Overbooking to Improve Patient
Access and Reduce Patient Backlog
Linda Laganga, Vp Of Quality Systems, Mental Health Center of
Denver, 4141 East Dickenson Place, Denver, CO, 80302, United
States of America, linda.laganga@mhcd.org, Stephen Lawrence
Joint Session DM/QSR: Quality and Statistical
Decision Making in Health Care Applications
Patient no-shows continue to trouble outpatient clinical service delivery. We
continue our piloting and implementation of scheduling models developed in our
earlier research to develop new techniques to assist clinics in meeting their goals
to improve patient flow and reduce backlog in scheduling. We utilize medical
practice experience to develop realistic estimates of costs and their effect on the
selection of high-performing scheduling alternatives.
Sponsored Session
Chair: Shuai Huang, University of Washington, Dept. of Industrial
and Systems Eng., Seattle, WA, United States of America,
shuai.huang.ie@gmail.com
1 - Social Media Analytics for the Promotion of Mental Health
Qingpeng Zhang, Assistant Professor,
City University of Hong Kong, Kowloon, Hong Kong - PRC,
qingpeng.zhang@cityu.edu.hk
3 - Improving HIV Early Infant Diagnosis Supply Chains in
Sub-Saharan Africa: Models and Application to Mozambique
Jonas Jonasson,Student, London Business School, Regent’s Park,
London NW1 4SA, United Kingdom, jjonasson@london.edu,
Sarang Deo, Jérémie Gallien
The digital footprints of Web users left on social media present important mental
health proxies. In this work, we aim to characterize the dynamics of the online
social groups for the mutual help of people suffering from depression. We
identified unique features in both language and social interaction patterns, and
interesting relationship between the two, which could have important
implications of the causes and factors of depression.
Most countries in sub-Saharan Africa experience delays in HIV early infant
diagnosis (EID). We develop a two-part modeling framework to generate
operational improvements in EID networks and evaluate their impact on public
health. For the case of Mozambique, we estimate that the interventions of
optimally re-assigning clinics to labs and optimally re-allocating diagnostic
capacity would result in 11% and 22% shorter turnaround times and 4% and 7%
more infants starting treatment, respectively.
2 - Adaptive Cluster-based Oversampling Method: Application to
Gynecological Surgery Failure Prediction
Iman Nekooeimehr, PhD Candidate, University of South Florida,
14304 Wedgewood Ct., Apt. 201, Tampa, FL, 33613, United
States of America, nekooeimehr@mail.usf.edu, Stuart Hart,
Allison Wyman, Susana Lai-yuen
■ TC30
A new oversampling method called Adaptive Semi-Unsupervised Weighted
Oversampling is presented for imbalanced dataset classification. It is adaptive, and
avoids overgeneralization and overfitting. The method was used with Support
Vector Machines to predict surgical failure after gynecological repair operations.
Results show 76% weighted accuracy and improvement over other oversampling
methods.
Decision Support Systems I
Contributed Session
Chair: Mohamad Hasan, Associate Professor, Kuwait University,
Department of Quantitative Methods &IS, CBA,Kuwait University,
Kuwait City, 13055, Kuwait, mkamal@cba.edu.kw
1 - Review of Consistency Among Pairwise Comparisons:
Relationship Between Indices and Human Perception
Yuji Sato, Graduate School of Management, Chukyo University,
101 Yagotohonmachi, Showa, Nagoya, 466-8666, Japan,
ysatoh@1988.jukuin.keio.ac.jp
3 - Real-time Detection of System Change Points via Graph Theoretic
Sensor Fusion
Prahalad Rao, SUNY Binghamton, 4400 Vestal Pkwy. E,
prao@binghamton.edu, Chou-An Chou, Samie Tootooni
We propose a novel graph theoretic approach for detection of system change
points from multidimensional sensor data. The approach is based on transform
time series data into an un-weighted and undirected planar graph, and
subsequently extracting topological invariants. This approach outperforms
conventional statistics-based monitoring techniques. We demonstrate the
effectiveness of the approach based on experimentally acquired sensor data from
advanced manufacturing processes and healthcare.
This paper reviews the Consistency Index (CI) of AHP. Since AHP requires
redundant pairwise comparisons, transitivity in judgment is often violated. The
review focuses on the detection capability of CI, and the relationship between the
size of CI and the goodness-of-fit of weight to decision maker’s perception. The
results imply that CI may not distinguish the consistency of judgment nor the size
may have no relation with the degree of goodness-of-fit of weight to decision
maker’s perception.
4 - High-throughput Screening for Rule Discovery from Highdimensional Datasets
Mona Haghighi, University of South Florida, 3202 E Fowler
Avenue, Tampa, FL, 33620, United States of America,
monahaghighi@mail.usf.edu, Shuai Huang, Xiaoning Qian,
Bo Zeng
2 - Transformations and Materializations of Uncertainty Sets in
Robust Optimization
Abhilasha Aswal, International Institute of Information
Technology, Bangalore, 26/C Electronics City, Bangalore, KA,
560100, India, abhilasha.aswal@iiitb.ac.in, Prasanna Gns
We propose a rule-based methodology to Identify risk-predictive baseline patterns
of Alzheimer’s disease through a network-based mathematical model. We apply
data-mining techniques to reduce dimensionality while taking care of synergistic
interaction of variables. Selecting a set of rules to monitor the progression of the
disease is the second part of this study.
We present a polyhedral representation of uncertainty for robust optimization and
a volume based uncertainty measure for it. Our decision support framework
enables easy transformations and materializations of a given uncertainty set and
also easy set-theoretic operations on alternative uncertainty sets. These operations
are quite useful in practice and are more difficult with probabilistic
representations of uncertainty and non-polyhedral robust uncertainty sets.
3 - Open Source or Proprietary? A Study on Software Diffusion in a
Competitive Market
Chao Ding, Assistant Professor, University of Hong Kong, KKL
807, Hong Kong, Hong Kong - PRC, chao.ding@hku.hk
■ TC32
Decision Support Systems for Data Mining
When choosing between open source software and proprietary software,
consumers will consider software quality, cost, consumer reviews, promotions,
compatibility, technical support, ease of use, etc. In this paper, we consider three
important decision making factors as identified in literature: external influence,
internal influence and ownership cost and study their impact on consumers’
adoption decision.
Contributed Session
Chair: Zhiguo Zhu, Associate Prof., Dongbei University of Finance and
Economics, No. 217 JianShan St., Shahekou District, Dalian, 110625,
China, zhuzg0628@126.com
1 - A Mixture Method of Multivariate Time Series Clustering
Cheng-Bang Chen, Penn State University, 233 Leonhard
czc184@psu.edu, Soundar Kumara
4 - A Decision Support System for Predicting International Freight
Flows for Trade
Mohamad Hasan, Associate Professor, Kuwait University,
Department of Quantitative Methods &IS, CBA, Kuwait City,
13055, Kuwait, mkamal@cba.edu.kw
Time series clustering is widely used in different domains. Although much
literature is available on time series clustering, only a few articles relate to
multivariate time series clustering. This research developed a clustering
methodology and applies different similarity/dissimilarity measures to
multivariate time series datasets. It can reduce the data size and has good
clustering performance.
A decision support system is developed that can help decision makers to take right
decisions about the country international trade system. It helps them to evaluate
deferent scenarios to improve the multimodal Transport system and import,
export, re-export, and transit operations. These improvements will enhance the
competitiveness and integration of this system.The overall results will help in
increasing the international trade share for the country.
328
TC34
2 - Mining Association between Promotions and Transactions to
Find Optimum Time for Targeted Promotions
Hari Koduvely, Dr, Samsung, # 2870,Orion Building, Outer Ring
Road, Bangalore, K, 560037, India, hari.koduvely@gmail.com,
Roshni Mohandas
types of new as well as returning patients and the available appointment slots, we
develop stochastic models to determine the number of slots and the scheduling
policy that optimize performance. Heuristic polices are proposed which share the
same structural properties of the optimal policy and are more computationally
efficient.
In the usual Targeted Promotion scenario one optimizes the promotion content
tailored towards a consumer’s purchase interests to maximize response. However
time of targeting a promotion also equally important. In this paper we present a
new method, based on the temporal association patterns between promotion and
transaction events, to find the optimum time for targeted promotions. We validate
our method against existing approaches on a real retail data set and show
significantly better results.
4 - Managing Virtual Appointments in Chronic Care
Armagan Bayram, Northwestern University, 2145 Sheridan Road,
Evanston, IL, 60201, United States of America,
abayram@northwestern.edu, Seyed Iravani, Sarang Deo,
Karen Smilowitz
Virtual visits can assist in managing chronic conditions by providing low cost
monitoring, treatment and education. Motivated by these benefits, we develop
capacity allocation models that decide which patients to schedule given limited
availability of both office and virtual visit slots. We model this problem using a
dynamic programming framework over a finite horizon, and perform analytical
and numerical analyses to identify policies for scheduling patients for different
medical interventions.
3 - The Effect of Non-local Diversity in Dynamic Class Prediction
Senay Yasar Saglam, PhD Student, University of Iowa, 108
Pappajohn Business Building, S210, Iowa City, IA, 52242, United
States of America, senayyasarsaglam@gmail.com, Nick Street
Classifiers’ agreement in the region where a new data instance resides in has been
considered a major factor in dynamic ensembles. We hypothesize that in this
region the agreement among classifiers that are different is more important than
among the similar ones. In other words, high local accuracy and confidence, and
high diversity in other regions, is desirable. In this study, we check the validity of
this hypothesis and verify that diversity still plays a role in the dynamic class
prediction.
■ TC34
Optimal Cancer Therapy
4 - Measuring Influence in Online Social Network Based on the
User-content Bipartite Graph
Zhiguo Zhu, Associate Prof., Dongbei University of Finance and
Economics, No. 217 JianShan St., Shahekou District, Dalian,
110625, China, zhuzg0628@126.com
Sponsored Session
Chair: Kevin Leder, Assistant Professor, University of Minnesota,
111 Church St, Minneapolis, MN, 55455, United States of America,
lede0024@umn.edu
1 - Nonstationary Spatiotemporally Integrated Fractionation
Ali Ajdari, University of Washington, Industrial and Systems
Engineering, Seattle, WA, 98195, ali.adr86@gmail.com,
Archis Ghate
How to precisely identify and measure influence has been a hot research
direction. Differentiating from existing researches, we are devoted to combining
the status of users in the network and the contents generated from these users to
synthetically measure the influence diffusion. In this paper, we firstly proposed a
directed user-content bipartite graph model. Finally, the experiment results verify
our proposed model can discover most influential users and popular broads
effectively.
We consider the optimal fractionation problem where the fluence-maps are
allowed to change over treatment sessions. This results in a high-dimensional
nonconvex dynamic optimization problem. We present an approximate solution
method rooted in convex and dynamic programming.
■ TC33
2 - Treatment of Chronic Myeloid Leukemia with Multiple
Targeted Therapies
Qie He, University of Minnesota, 111 Church Street SE,
Minneapolis, MN, United States of America, qhe@umn.edu,
Junfeng Zhu, Kevin Leder, Jasmine Foo
Appointment Scheduling in Healthcare
Sponsored Session
Recently several targeted therapies have been developed to treat Chronic Myeloid
Leukemia (CML). A significant problem is the development of resistance to
therapy in patients. Therapy combination can slow this development, but the
number of combinations is huge. We develop a model to find combinations that
are promising for clinical trials. The model captures cell evolution and toxicity
constraints. Our optimal combinations are predicted to significantly outperform
common clinical practice.
Chair: Armagan Bayram, Northwestern University, 2145 Sheridan
Road, Evanston, IL, 60201, United States of America,
abayram@northwestern.edu
1 - Ensuring Timely Access and Adequate Capacity for an
Endocrinology Clinic
Moses Chan, University of Michigan, 1205 Beal Avenue,
Ann Arbor, MI, United States of America, mosesyhc@umich.edu,
Amy Cohn, Amy Rothberg
3 - Combined Therapy in Acute Lymphoblastic Leukemia
Kevin Leder, Assistant Professor, University of Minnesota, 111
Church St, Minneapolis, MN, 55455, United States of America,
lede0024@umn.edu
The weight management program was designed to promote weight loss for
morbidly obese patients. Program participation is associated with reductions in
BMI and improvements in cardiovascular risk factors and quality of life. Providers
are booked weeks out, posing a challenge to schedule consecutive weekly visits.
Non-adherence to schedule undermines the effectiveness of the program. The
purpose of this study is to improve patient compliance with the program and to
increase program access.
Acute lymphoblastic leukemia is a cancer of the blood system. While successful
treatment of the disease in juvenile patients is possible, it is difficult to treat in
adult patients. One treatment modality is the use of the targeted therapy nilotinib.
However, drug resistance is a serious issue. To avoid this drug resistance we
consider the combination of nilotinib and radiation. We develop a mathematical
model for this combined therapy and compare with experimental observations.
2 - An Online Appointment Scheduling Model
Ali Kemal Dogru, Om PhD Student, University of Alabama, 315
Bidgood Hall, 361 Stadium Drive, Tuscaloosa, AL, 35487, United
States of America, akdogru@crimson.ua.edu, Sharif Melouk
Incorporating patient centered medical home (PCMH) principles, we develop an
online appointment scheduling system (OASS) for a primary care setting. We
propose a simulation optimization solution approach that uses two models
working in concert to provide high quality solutions (i.e., schedules) in short
time. We aim to minimize: 1) weighted cost of expected patient waiting time and
2) doctor idle time and overtime. Key Words: Online Appointment Scheduling,
Simulation Optimization, PCMH
3 - Managing Series Patients in a Healthcare Facility
Siyun Yu, STOR Department, UNC-Chapel Hill, B26 Hanes Hall,
yusiyun@live.unc.edu, Vidyadhar Kulkarni, Vinayak Deshpande
Series patients are scheduled for a series of appointments, such as patients in
physical therapy clinic, dialysis center, etc. To balance the demands from different
329
TC35
■ TC35
■ TC36
Disaster and Emergency Management I
Humanitarian Applications III
Contributed Session
Sponsored Session
Chair: Rafael Diaz, Research Associate Professor, Old Dominion
University, 1040 University Blvd, Suffolk, VA, 23435,
United States of America, rdiaz@odu.edu
1 - An Optimization Model for Seismic Hazard Loss Analysis for
Spatially Distributed Infrastructure
Hasan Manzour, Industrial & Systems Engineering, University of
Washington, Box 352650, Seattle, WA, 98195-2650, United
States of America, hmanzour@uw.edu, Rachel Davidson
Chair: Melih Celik, Middle East Technical University, ODTÜ Kampüsü
Endüstri Mühendisligi, Oda 219 Cankaya, Ankara, 06800, Turkey,
cmelih@metu.edu.tr
1 - Pre-disaster Unmanned Air Vehicle Base Location and Routing for
Road Damage Assessment and Repair
Seyyed Kian Farajkhah, Middle East Technical University, METUCankaya, Metu Campus Endöstri Möhendisligi, Ankara, 06800,
Turkey, kian.farajkhah@metu.edu.tr, Melih Celik
The new Optimization-based Probabilistic Scenario method produces a small set
of probabilistic ground motion maps to represent the seismic hazard for analysis
of spatial distributed infrastructure. A set of just 124 ground motion maps were
able to match the hazard curves based on a million-year Monte Carlo simulation.
This enormous computational savings has substantial implications for regionalscale since it can allow many more downstream analyses.
Following large-scale disasters, unmanned air vehicles (UAVs) can help efficiently
gather data on the status of the roads in the network. Given a set of potential
disaster scenarios, we address the problem of establishing connectivity between
relief supply and demand by means of road repair. A two-stage stochastic model is
developed to determine a UAV base location and time-limited routes so that the
expected shortest path length between the supply and demand nodes is
minimized.
2 - A Model of the Effect of Pandemic Influenza on the
U.S. Blood Supply
Hussein Ezzeldin, FDA, CBER, 10903 New Hampshire Ave,
Bldg 71 Rm 1009C, Silver Spring, MD, 20993, United States of
America, hussein.ezzeldin@fda.hhs.gov, Arianna Simonetti,
Richard Forshee
2 - Disaster Operations Management: Recovery Classification and
Research Framework
Niratcha Grace Tungtisanont, PhD Candidate, Clemson
University, 100 Regency Dr, #22, Central, SC, 29630,
United States of America, ntungti@g.clemson.edu, Aleda Roth,
Yann Ferrand
We present the spatial and temporal impact of Pandemic Influenza (PI) on the US
blood supply through an inter-regional blood transfer system. We utilize a hybrid
optimization heuristic to enhance the global performance of the network. Using
Neural Networks trained by Particle Swarm Optimization, we model a function of
regional factors to optimize the daily blood transfers among US regions. We
simulate the effect of PI on regional blood transfers and compare to those during
normal operations.
We propose a research framework for improving post-disaster phase recovery. We
address what types of investments should be made and their relative allocations
in the “pre” and “during” emergency phases to improve the effectiveness of the
recovery process? We use the proposed framework to draw managerial and policy
implications.
3 - Agent-Based Modeling to Simulate Resilience of Water Systems
for Healthy and Secure Communities
Emily Berglund, Associate Professor, North Carolina State
University, CB 7908, Raleigh, NC, 27695, United States of
America, emily_berglund@ncsu.edu, Jacob Monroe, Hayden
Strickling, Michael Knepper, Elizabeth Ramsey, M. Ehsan Shafiee
■ TC37
Kidney Allocation and Exchange
Civic water systems are vulnerable to attacks and disasters that threaten the
health and security of communities. When water service is lost due to a water
quality failure or an attack on critical infrastructure, the decision-making of
perpetrators, security personnel, utility managers, and the public can influence
event outcomes. An agent-based modeling approach is developed to simulate the
impact of sensing, communication, security, and infrastructure management on
community resilience.
Contributed Session
Chair: Naoru Koizumi, Assoc Professor, GMU, 3351 N Fairfax Dr,
Arlington, VA, 22203, United States of America, nkoizumi@gmu.edu
1 - The Dynamics of Kidney Exchange
John Dickerson, CMU, 9219 Gates-Hillman Center, Pittsburgh,
PA, 15213, United States of America, dickerson@cs.cmu.edu,
Tuomas Sandholm
4 - Decision Support to Air Rescue Unit Allocation in Disaster
Management Operations
Sergio Reboucas, ITA, Rua H9C, Apt. 302, São Jose Dos Campos,
SP, 12228612, Brazil, reb@ita.br
We discuss analytic, optimization, and game-theoretic approaches to matching in
dynamic kidney exchange. We consider dynamism (i) at the post-match pretransplant stage (ii) as patients and donors arrive and depart over time, and (iii)
as multiple exchanges compete for overlapping sets of participants. We
empirically validate our models and theoretical results on over 150 match runs of
the UNOS national kidney exchange.
After a disaster break up, rescue helicopters have a valuable role in response
phase. The allocation of these air rescue units requires knowledge about certain
conditions that are most of times uncertain and its analysis and trade-offs must be
thoroughly done. This paper aims to suggest a methodology to support the air
rescue unit allocation decision in a natural disaster response phase context.
2 - A New Model to Decide Kidney–Offer Admissibility Dependent on
Patients’ Lifetime Failure Rate
Michael Bendersky, Ben Gurion University of the Negev,
Beersheba, Israel, michael.bendersky@gmail.com, Israel David
5 - Modeling Housing Stock Recovery after a Catastrophic
Storm Event
Rafael Diaz, Research Associate Professor, Old Dominion
University, 1040 University Blvd, Suffolk, VA, 23435, United
States of America, rdiaz@odu.edu, David Earnest, ManWo Ng,
Joshua G. Behr
We propose a new model to decide kidney-offer admissibility depending on
patient’s age, estimated lifetime probabilistic profile and prospects on the waiting
list. We allow for a broad family of lifetime distributions - Gamma - thus enabling
flexible modeling of one’s survival under dialysis. It yields the optimal critical
times for acceptance of offers of different qualities and may serve the organizer of
a donation program, the surgeon and the individual recipient practicing patientchoice.
Severe catastrophic storm events adversely affect housing stock and regional
capacity to build and repair houses. Rebuilding this capacity takes time while the
region faces an unexpected surge in the demand. We present a simulation model
that considers a supply chain perspective. The model provides significant insights
for policy makers into how the production of permanent housing depends upon
the uncertainties and feedback effects of material, labor, funds, and regulatory
environments.
3 - Preemptive Approach to Kidney Allocation in USA
Philip Appiahk-Kubi, Ohio University, 14 Pine St, Apt. #1B, The
Plains, OH, 45780, United States of America, pa809911@ohio.edu
The new kidney allocation policy improves kidney utilization. However, the
policy has no consideration for allocation of cadaveric kidneys under emergency
situations; a problem observed by the National Kidney Foundation. This research
evaluates a point scoring model with considerations for emergency allocation.
Simulated results indicate that the model minimizes number of waitlist deaths by
2% while prioritizing sensitive candidates and waiting time.
330
TC40
■ TC39
4 - Optimal Integration of Kidney Exchange Programs with Antibody
Reduction Therapy
Naoru Koizumi, Assoc Professor, GMU, 3351 N Fairfax Dr,
Arlington, VA, 22203, United States of America,
nkoizumi@gmu.edu, Monica Gentili, Keith Melancon
39-Room 100, CC
Distribution Channel Management
Invited Session
Kidney paired donation (KPD) allows incompatible pairs to exchange kidneys
with other incompatible pairs. However, evidence suggests there stills exist
barriers to KPD utilization, especially among difficult-to-match transplant
candidates and positive actual or virtual crossmatches. Using mathematical
models, we investigate how to optimally integrate antibody reduction therapy in
KPD to increase successful living-donor kidney transplants among difficult to
match candidates.
Chair: Xiaowei Xu, Associate Professor, Rutgers Business School-New
Brunswick, 100 Rockafeller Rd., Piscataway, NJ, 08854,
United States of America, xiaoweix@andromeda.rutgers.edu
1 - Co-Advertising and Channel Power in Distribution Channels
Xiaowei Xu, Associate Professor, Rutgers Business School-New
Brunswick, 100 Rockafeller Rd., Piscataway, NJ, 08854,
United States of America, xiaoweix@andromeda.rutgers.edu
■ TC38
We study a manufacturer-retailer channel, in which the manufacturer decides
wholesale price as the channel leader and the retailer decides the retail price as
the channel follower. Besides the retail price, customer demand is influenced by
non-price marketing instruments, such as advertising. We identify business
scenarios, under which the manufacturer should not participate any coadvertising campaign even if it’s free, since doing so will increase the channel
power of the retailer.
Queueing Models II
Contributed Session
Chair: Benjamin Legros, Ecole Centrale Paris,
Grande Voie des Vignes, Chatenay-Malabry, 92290, France
1 - A New Look at Markov Processes of G/M/1-type
Jason Joyner, PhD Student, Clemson University, Clemson
University O-110 Martin Hall, Box 340975, Clemson, SC, 29634,
United States of America, jjoyner@g.clemson.edu
2 - Effects of Channel Intermediaries on Quality-price Competition
S. Chan Choi, Rutgers Business School-New Brunswick,
100 Rockafeller Road, Piscataway, NJ, 08854,
United States of America, chanchoi@rci.rutgers.edu
We show that when products are vertically differentiated, the optimal channel
structure depends on whether a company is a high- or low-quality producer.
Either manufacturer benefits by channel integration while the competitor uses an
intermediary, but this effect is stronger for the low-quality manufacturer. If an
intermediary is to be used, the low-quality manufacturer has more incentive to
use an exclusive dealer. But the total channel profit is higher with a common
retailer.
We present a new method for deriving the stationary distribution of an ergodic
Markov process of G/M/1-type in continuous-time. Our method derives and
makes use of a new representation for each element of the rate matrices
contained in the stationary distribution. This method can also be modified to
derive the Laplace transform of each transition function associated with Markov
processes of G/M/1-type.
2 - Exact Simulation of Non-Stationary Reflect Brownian Motion
Mohammad Mousavi, Assistant Professor, University of
Pittsburgh, 1048 Benedum Hall, Pittsburgh, PA, 15212,
United States of America, mousavi@pitt.edu
3 - Fairness in Supply Chain Contracts with Sales Efforts
Ju Myung Song, Rutgers Business School, Room 430, 1
Washington Park, Newark, NJ, 07102, United States of America,
jumyung.song@rutgers.edu
We discuss the challenges that arise in the planning simulations of systems with
time dependent arrival and service rates. Estimating how far back in time a
simulation must be initialized is an essential problem in planning simulations. We
propound using reflected Brownian motion (RBM) with time-dependent drift and
volatility as a guide for estimating this initialization time. We develop the first
exact simulation method for RBM with time-dependent drift and volatility.
Fairness is an important incentive for supply chain contract design. I consider a
setting where a retailer chooses both retail price and sale effort to maximize
profit, and analyze how fairness in a supply chain affects supplier and retailer’s
behaviors and their expected profits.
3 - Traffic Volume and Travel Time Variability under
Random Interruptions
Marcelo Figueroa, PhD Student, Rutgers University, 93 Marvin
Lane, Piscataway, NJ, 08854, United States of America,
marcelo.figueroa@rutgers.edu, Melike Baykal-görsoy
4 - A Longitudinal Analysis of Supplier Working Relations in
Component Markets
Sengun Yeniyurt, Associate Professor, Rutgers Business School,
100 Rockafeller Rd, Piscataway, NJ, 08854, United States of
America, yeniyurt@business.rutgers.edu, Steven Carnovale,
John W. Henke
We show the advantages of modeling the number of vehicles on a freeway
corridor as an M/M/Infinity queueing system subject to random service
degradation in order to obtain variability estimates for congestion and travel time
delay. We make use of the analytical stationary distribution of the number of
customers thus avoiding the use of traffic simulation. We validate our approach
by using traffic count data, and relevant weather events and traffic incidents as
causes of service degradation.
This study utilizes a longitudinal dataset that includes information regarding
supplier working relations and sourcing transactions in the North American
Automotive industry. Econometric models are developed and estimated utilizing
information provided by first tier component suppliers to major automotive
manufacturers. The estimates reveal the interplay between past interactions,
future expectations, and working relations and their effect on transactional
decisions in component markets.
4 - Admission Control Policies for Multi-Channel Call Centers:
Should We Delay the Call Rejection?
Benjamin Legros, Ecole Centrale Paris, Grande Voie des Vignes,
Chatenay Malabry, France, belegros@laposte.net
■ TC40
40- Room 101, CC
We study strategies of rejection in call centers with inbound and outbound calls.
The firm is looking for the best possible trade-off between the inbound and
outbound calls performance. Rejection at arrivals, so-called rejection “a priori”
and rejection after experimenting some wait, so-called rejection “a posteriori” are
considered. Our main finding is that rejection a posteriori provides a better
performance in terms of waiting time for served customers than a rejection a
priori.
Marketing I
Contributed Session
Chair: Robert Bordley, Expert Systems Engr Professional,
Booz-Allen-Hamilton, 525 Choice Court, Troy, MI, 48085,
United States of America, Bordley_Robert@bah.com
1 - How Cultural Difference Influences Consumer Behavior in
Hypermarket Industry
Mei-Wen Chao, Assistant Professor, Kao Yuan University, 1821
Zhongshan Rd., Luzhu Dist., Kaohsiung, 82151, Taiwan - ROC,
t80149@cc.kyu.edu.tw
No empirical research exists to discuss the issues of culture and consumer
behavior in the hypermarket industry using the territory of Taiwan and the U.S.
as the units of comparison. This paper attempted to explore grocery consumers’
inner world and how their shopping perceptions vary between Taiwanese and
American cultures. The contexts of interviews are given and additional findings
are also put forward. Salient results and practical issues are discussed in detail in
this paper.
331
TC41
2 - The Effect of Price Ending on Consumer Behavior
Yoshiyuki Okuse, Professor, Senshu University, 2-1-1,
Higashi-Mita, Tama-Ku, Kawasaki, 2148580, Japan,
okuse@isc.senshu-u.ac.jp
3 - Dynamic New Patient Consult Scheduling for Medical Oncology
Antoine Sauré, University of British Columbia, 2053 Main Mall,
Vancouver, BC, V6T 1Z2, Canada, antoine.saure@sauder.ubc.ca,
Claire Ma, Jonathan Patrick, Martin Puterman
In the area of pricing research, a lot of researches on price endings have been
developed. The purpose of this research is to specify the effect of price endings on
consumer behavior in Japan.
Motivated by an increasing demand for cancer care and long waits for new
patient consults, we undertook a study of medical oncology scheduling practices
at a regional cancer center. As a result, we formulated and approximately solved a
discounted infinite-horizon MDP model that seeks to identify policies for
allocating oncologist consultation time to incoming new patients, while reducing
waits in a cost-effective manner. The benefits from the proposed method are
evaluated using simulation.
3 - I Like using My Mobile Apps But…
A Study of Post Consumption Appraisal
Anubha Mishra, Assistant Professor Of Marketing,
University of Idaho, 875 Perimeter Dr, Moscow, ID, 83844,
United States of America, amishra@uidaho.edu
4 - Optimal Issuing Policies for Hospital Blood Inventory
Alireza Sabouri, Assistant Professor, Haskayne School of Business,
University of Calgary, Calgary, AB, Canada,
alireza.sabouri@haskayne.ucalgary.ca, Steven Shechter, Tim Huh
The study of mobile app consumption suggested three distinct dimensions. Users’
evaluation of control, freedom, newness, assimilation, and fulfillment of need
from apps was captured by Perceived Benefits; Perceived Apprehension, covered
chaos, enslavement, obsolesce, isolation, and creation of needs and Perceived
Obscurity, examined ambiguity. Perceived usefulness positively influenced all
dimensions indicating that while apps may be perceived as helpful, it can also
create isolation.
We propose a model for allocating red blood cells for transfusion to patients,
which is motivated by recent evidence suggesting that transfusing older blood is
associated with increased mortality rate. We study the properties of blood
issuance policies that balance the trade-off between “quality” measured in
average age of blood transfused and “efficiency” measured in the amount of
shortage. Based on our analysis, we design efficient issuance policies and evaluate
their performance.
4 - Increasing User Engagement with Mobile Analytics
Chaitanya Sagar, Chief Executive Officer, Perceptive Analytics,
353 West 48th Street, New York, NY, 10036,
United States of America, cs@perceptive-analytics.com
Mobile represents a tectonic platform shift with great opportunities and
challenges. 80% of users do not return to an app after the first day of
downloading it. 80% of total app revenue is ‘in-app’ purchases - so unless an app
can engage users, it cannot generate significant revenue. Add to that, top 20%
apps generate 97% of the revenue making fierce competition among 1.2 million
apps. I focus on the heart of this problem increasing engagement with users
specifically using push-notifications.
■ TC42
42-Room 102B, CC
Joint Session MSOM-Health/HAS: Workarounds,
Errors and Interruptions in Healthcare
Sponsored Session
5 - Maximum Entropy Models of Individual Choice
Robert Bordley, Expert Systems Engr Professional, Booz-AllenHamilton, 525 Choice Court, Troy, MI, 48085, United States of
America, Bordley_Robert@bah.com, Ehsan Soofi
Chair: Anita Tucker, Associate Professor, Brandeis University,
415 South Street, Waltham, MA, 02453, United States of America,
atucker@brandeis.edu
1 - Hospital Operations and Patient Satisfaction
Sarah Zheng, Doctoral Candidate, Boston University, 16 Gold Star
Rd., Cambridge, MA, 02140, United States of America,
xinzheng@bu.edu, Amy McLaughlin, Aubrey Podell,
Anita Tucker, Z. Justin Ren
Many forecasts are based on economic models of individual choice. But these
models assume actual individual choice is rational, an assumption which some
viewed as having been refuted. To avoid making this assumption, this paper
shows that maximum entropy models can approximate general discrete choice
models. This paper also shows how to parameterize such models in order to use
them for forecasting.
We look at the impact of operations performance on service quality. Our study
site is a nationally-ranked major hospital in the Boston area. Service quality is
measured by both medical errors and patient satisfaction. Daily operations are
measured by performance in about a dozen of its supporting services. We attempt
to answer questions such as: What are the operational drivers of medical errors?
To what extent does higher operations performance lead to higher patient
satisfaction?
■ TC41
41-Room 102A, CC
Joint Session MSOM-Health/HAS: Healthcare
Operations
Sponsored Session
2 - Medical Errors in the Healthcare Delivery: An Econometric
Analysis of the Operational Sources
Sriram Thirumalai, Associate Professor, Texas Christian
University, Neeley School of Business, TCU Box 298530,
Fort Worth, TX, 76116, United States of America,
s.thirumalai@tcu.edu
Chair: Alireza Sabouri, Assistant Professor, Haskayne School of
Business, University of Calgary, Calgary, AB, Canada,
alireza.sabouri@haskayne.ucalgary.ca
1 - A Queueing Model for Liver Transplant Waiting List Process
Zinan Yi, Operations Research, North Carolina State University,
Raleigh, NC, United States of America, zyi@ncsu.edu, Maria
Mayorga, Stephanie Wheeler, Sidney Barritt, Eric Orman
Medical errors in the delivery of care is a significant cause of concern in
healthcare supply chains. Based on an econometric analysis of a panel dataset on
medical errors, this study serves to examine the sources of medical errors and
error mitigation in the delivery of care in hospitals.
3 - The Impact of Workarounds on Patient Falls and Pressure Ulcers
Anita Tucker, Associate Professor, Brandeis University, 415 South
Street, Waltham, MA, 02453, United States of America,
atucker@brandeis.edu
Liver transplant is the only therapy for patients with end stage liver disease. The
composition and dynamics of the waiting list are the interest of both patients and
doctors. In this paper, we used the United Network for Organ Sharing and Organ
Procurement and Transplantation Network database to develop a queueing model
for the waiting list population. Using the model, we will predict future waiting list
and other characteristics.
We present results from a survey of 100 medical/surgical nursing units that tests
the impact of workarounds and operational failures on nursing sensitive patient
outcomes, such as pressure ulcers, falls, patient satisfaction and infections.
2 - Investigating Steroid Withdrawal Strategies for Kidney
Transplant Recipients
Yann Ferrand, Assistant Professor, Clemson University,
100 Sirrine Hall, Clemson, SC, 29634, United States of America,
yferran@clemson.edu, Vibha Desai, Christina Kelton,
Teresa Cavanaugh, Jaime Caro, Jens Goeble, Pamela Heaton
4 - Batching of CSP Medication in In-Hospital Pharmacy
Vera Tilson, Simon School of Business, University of Rochester,
Rochester, NY, 14627, United States of America,
vera.tilson@simon.rochester.edu, Gregory Dobson, David Tilson
Hospital pharmacy departments batch production of Compounded Sterile
Products (CSP). A change in a patient’s condition can lead to change or
cancellation of physician’s orders. A very large proportion of orders are cancelled,
which leads to waste custom compounded medication. We create an integer
programming model to help pharmacies plan batch production trading off the cost
of waste and the cost of employee labor.
We evaluate various steroid withdrawal strategies for kidney transplant recipients.
The goal is to minimize major complications resulting from these complex drug
regimens over the long term. We develop a model calibrated with an econometric
study of patient data from a national registry to simulate the long-term course of
these patients. We report on the frequency and timing of adverse events and
identify trade-offs in the steroid withdrawal strategies.
332
■ TC43
TC45
2 - Mobile Technology in Retail: The Value of
Location-based Information
Marcel Goic, Assistant Professor Or Marketing, University of
Chile, Republica #701, Santiago, 8370438, Chile,
mgoic@dii.uchile.cl, Jose Guajardo
43-Room 103A, CC
Joint Session RMP/MSOM: Choice Models:
Estimation and Optimization
We analyze the value of location-based information in mobile retailing and the
conditions under which incorporating geolocation information increase
effectiveness metrics for retailers.
Sponsored Session
3 - Dynamic Pricing in Ride-Sharing Platforms
Siddhartha Banerjee, Postdoc, Stanford University, 475 Via
Ortega, Stanford, CA, 94305, United States of America,
sidb@stanford.edu, Carlos Riquelme, Ramesh Johari
Chair: Sumit Kunnumkal, Indian School of Business, Gachibowli,
Hyderabad, India, Sumit_Kunnumkal@isb.edu
1 - Formulation, Motivation, and Estimation for the D-Level Nested
Logit Model
Guang Li, University of Southern California, Bridge Hall 401,
None, Los Angeles, CA, 90089-0809, United States of America,
guangli@usc.edu, Huseyin Topaloglu, Paat Rusmevichientong
We develop a model for ride-share platforms, which combines a queueing model
for the platform dynamics with strategic models for passenger and driver
behavior. Using this, we study various aspects of this system - the value of
dynamic pricing versus static pricing; the robustness of these policies; the effect of
heterogenous ride-request rates and traffic between different locations. Joint
work with Ramesh Johari, Carlos Riquelme and the Data Science team at Lyft.
Using a tree of depth d, we provide a novel formulation for the d-level nested
logit model. Our model is consistent with the random utility maximization
principle and equivalent to the elimination by aspects model. Using new
concavity results on the log-likelihood function, we develop an effective
parameter estimation algorithm. Numerical results show that the prediction
accuracy of the d-level nested logit model can be substantially improved by
increasing the number of levels d in the tree.
4 - Pricing with Limited Knowledge of Demand
Maxime Cohen, MIT, 70 Pacific Street, Apt. 737B, Cambridge,
MA, 02139, United States of America, maxcohen@mit.edu,
Georgia Perakis, Robert Pindyck
How should a firm price a new product with limited information on demand? We
propose a simple pricing rule that can be used if the firm’s marginal cost is
constant: the firm estimates the maximum price it can charge and then sets price
as if demand were linear. We develop bounds that show that if the true demand is
one of many commonly used demand functions, the firm will do “very well” - its
profit will be close to what it would earn if it knew the true demand.
2 - Assortment Optimization Over Time
James Davis, Cornell University, 290 Rhodes Hall, Ithaca, NY,
United States of America, jamesmariodavis@gmail.com,
Huseyin Topaloglu, David Williamson
Inspired by online retail we introduce a new type type of assortment optimization
problem: assortment optimization over time. In this problem the retailer must
choose which products to display but must also choose an ordering for the
products. This is a relevant problem when items are displayed as a list; this is
common when returning results from a search query, for example. We provide a
framework to analyze this problem, provide an approximation algorithm, and
some hardness results.
■ TC45
45-Room 103C, CC
Behavioral Issues in RM
3 - Tractable Bounds for Assortment Planning with Product Costs
Sumit Kunnumkal, Indian School of Business, Gachibowli,
Hyderabad, India, Sumit_Kunnumkal@isb.edu,
Victor Martínez-de-Albéniz
Sponsored Session
Chair: Anton Ovchinnikov, Queen’s University, 143 Union Str West,
Kingston, Canada, anton.ovchinnikov@queensu.ca
1 - Should Consumers be Strategic?
Arian Aflaki, Doctoral Student, Duke University, 100 Fuqua
Drive, Box 90120, Durham, NC, 27708, United States of America,
arian.aflaki@duke.edu, Robert Swinney, Pnina Feldman
Assortment planning under a logit demand model is a difficult problem when
there are product specific costs associated with including products into the
assortment. In this paper, we describe a tractable method to obtain an upper
bound on the optimal expected profit. We provide performance guarantees on the
upper bound obtained. We describe how the method can be extended to
incorporate additional constraints on the assortment or multiple customer
segments.
We consider whether strategic consumer behavior benefits consumers when they
purchase from a rational, revenue-maximizing firm that sets prices over multiple
periods. We show that strategic behavior does not benefit all consumers. Then, by
studying a wide range of pricing and inventory strategies in a unified setting, we
find that different strategies may induce different levels of interest in strategic
behavior.
4 - Clustering Consumers Based on Their Preferences
Ashwin Venkataraman, New York University,
715 Broadway, New York, NY, United States of America,
ashwin.venkataraman@gmail.com, Srikanth Jagabathula,
Lakshminarayana Subramanian
2 - Intertemporal Pricing under Minimax Regret
America, yliu2@stern.nyu.edu, Rene Caldentey, Ilan Lobel
Preference-based clustering is an important and challenging problem. We propose
a non-parametric method to cluster consumers based on their preferences for a
set of items. Our method combines the versatility of model-free clustering (such
as k-means) with the flexibility and rigor of model-based clustering (based on EM
algorithm). Our approach is fast, can handle missing data, identify general
correlation patterns in consumer preferences, and has provable guarantees under
reasonable assumptions.
We consider a monopolist selling a product to a population of consumers who are
heterogeneous in valuations and arrival times. We study the policies that attain
minimum regret when selling to either myopic or strategic customers. We
characterize the set of optimal policies and demonstrate their structural
properties.
■ TC44
3 - Behavioral Anomalies in Consumer Wait-or-Buy Decisions and the
Implications for Markdown Management
Nikolay Osadchiy, Emory University,
1300 Clifton Rd NE, Atlanta, GA, 30322, United States of
America, nikolay.osadchiy@emory.edu, Anton Ovchinnikov,
Manel Baucells
44-Room 103B, CC
Pricing in Online Markets
Sponsored Session
A decision to buy at a tag price or wait for a possible markdown involves a tradeoff between the value, delay, risk and markdown magnitude. We build an
axiomatic framework that accounts for three well-known behavioral anomalies
along these dimensions and produces a parsimonious generalization of discounted
expected utility. We consider a pricing/purchasing game and show that
accounting for the behavioral anomalies results in substantially larger markdowns
and leads to noticeable revenue gains.
Chair: Kostas Bimpikis, Stanford GSB, 655 Knight Way, Stanford, CA,
94305, United States of America, kostasb@stanford.edu
1 - Modeling Growth for Services: Evidence from the App Economy
Ken Moon, PhD Candidate, Stanford GSB, 655 Knight Way,
Stanford, CA, 94305, United States of America,
kenmoon@stanford.edu, Haim Mendelson
We present a model of service operations to grow and sustain customers by the
operational design and performance of services, rather than marketing alone.
Applying our framework to data from services in the app economy, we show (i)
that customers’ engagement contributes as powerfully to growth as virality; and
(ii) evidence of an experience curve (from customer interactions) for service
operations. We present a model of incentive-compatible pricing for this setting.
333
TC46
4 - Selling a Dream: Pricing under Savoring and Anticipation
Javad Nasiry, Assistant Professor, Hong Kong University of
Science and Technology, ISOM, LSK Building, HKUST, Hong
Kong, Hong Kong - PRC, nasiry@ust.hk, Ioana Popescu
■ TC47
We study a market where customers derive emotional utility from anticipating
pleasurable purchase outcomes, but experience disappointment if outcomes fall
short of what they anticipated. In this context, we show that firms can profit by
adopting randomized pricing policies.
Operations
Sponsored Session
■ TC46
Chair: Gal Raz, Associate Professor, Ivey Business School, Western
University, 1255 Western Road, London, ON, Canada,
RazG@darden.virginia.edu
47-Room 104B, CC
Topics in Remanufacturing and Recycling
46-Room 104A, CC
Co-Chair: James Abbey, Texas A&M University, 4217 TAMU 320P,
jabbey@mays.tamu.edu
1 - Recycling as a Strategic Supply Source
Gal Raz, Associate Professor, Ivey Business School, Western
University, 1255 Western Road, London, ON, Canada,
RazG@darden.virginia.edu, Gilvan (Gil) Souza
Issues Related to Supply Chain Management
Sponsored Session
Chair: Achal Bassamboo, Professor, Kellogg School of Management,
2001 Sheridan Road, Evanston, IL, 60208, United States of America, abassamboo@kellogg.northwestern.edu
1 - Worker Poaching in a Supply Chain: Enemy from Within?
Evan Barlow, Northwestern University, Evanston, IL,
United States of America, e-barlow@kellogg.northwestern.edu,
Gad Allon, Achal Bassamboo
In this paper we investigate how recycling can be used as a strategic source of
supply in the presence of competition and a powerful material supplier. We
examine the economic and environmental impact of a manufacturer’s decision to
recycle its products and the implications on the customers, supplier and society as
a whole.
2 - The Effect of Environmental Regulation on DFE Innovation:
Social Cost in Primary/ Secondary Markets
Cheryl Druehl, George Mason University, 4400 University Dr MS
5F4, Fairfax, VA, 22030, United States of America,
cdruehl@gmu.edu, Vered Blass, Gal Raz
Poaching workers has become a universal practice. We explore worker poaching
between firms linked in a supply chain. We show that the classical intuition from
labor economics is insufficient in explaining poaching between supply chain
partners. We also show how and under what conditions worker poaching can
actually improve supply chain performance. Finally, we show how the
equilibrium identity of the supply chain bottleneck depends on the interaction
between hiring, poaching, and productivity.
We examine DfE innovations in the use stage and for refurbishing of a firm selling
new primary market products and refurbished products in a separate secondary
market. The firm determines innovations, prices, and fraction collected. Using
LCA data from cell phones, we compare EPR and Use stage regulations on profits
and environmental impact.
2 - Dynamic Clustering and Assortment Personalization:
The Value of Information Pooling
Sajad Modaresi, Duke University, 100 Fuqua Drive, Durham, NC,
United States of America, sajad.modaresi@duke.edu, Denis Saure,
Fernando Bernstein
3 - New Versus Refurbished: Key Factors that Influence
Consumers’ Decisions
Erin Mckie, University of South Carolina, 1014 Greene Street,
Columbia, SC, 29208, United States of America,
erinmckie@gmail.com, Mark Ferguson, Michael Galbreth,
Sriram Venkataraman
A retailer faces heterogeneous customers with initially unknown preferences. The
retailer can personalize assortment offerings based on available profile
information; however, users with different profiles may have similar preferences,
suggesting that the retailer can benefit from pooling information among
customers with similar preferences. We propose a dynamic clustering approach
that adaptively adjusts customer segments and personalizes the assortment
offerings to maximize cumulative profit.
Remanufacturing is increasingly providing new profit opportunities for firms, and
more product condition options – such as new, refurbished, and used – for
consumers to choose between. Using secondary data and choice model analysis
techniques, this study estimates the influence of various factors on consumers’
purchasing decisions.
3 - Policing a Self-policing Firm: Incentives for Detection and
Disclosure of Compliance Violations
Sang Kim, Yale School of Management, New Haven, CT,
United States of America, sang.kim@yale.edu
4 - The Value of Competition in Remanufacturing
Narendra Singh, Georgia Institute of Technology, Atlanta, GA,
United States of America, Narendra.Singh@scheller.gatech.edu,
Karthik Ramachandran, Ravi Subramanian
One of the challenges in enforcement of environmental regulations is designing
an effective incentive mechanism that elicits firms’ voluntary detection and
disclosure of compliance violations. with a right incentive, a firm self-polices its
internal operations to detect random violations before a regulator does, and
subsequently puts a remedial action in place. We study this incentive dynamic
using a game-theoretic framework.
We study an OEM’s product strategy when the OEM offers a new product that
depreciates over time and consumers are strategic. The OEM competes with a
third-party remanufacturer for acquisition and remanufacturing of the
depreciated products. We study how competition from the third-party
remanufacturer affects the OEM.
4 - Reshoring Manufacturing: Supply Availability, Demand Updating,
and Inventory Pooling
Bin Hu, Assistant Professor, UNC Kenan-Flagler Business School,
CB#3490 McColl Bldg, University of North Carolina,
Bin_Hu@kenan-flagler.unc.edu, Li Chen
■ TC48
48-Room 105A, CC
Managing Finances and Risk in Supply Chains
Reshoring shortens the distance from factory to market, however limited onshore
supply availability may force reshoring manufacturers to remain dependent on
offshore suppliers, leading to increased distance from supplier to factory. In this
case, we show that manufacturers’ preferences toward reshoring boil down to
trade-offs between operational flexibilities. We characterize when manufacturers
prefer reshoring, and further identify operational strategies that can swing such
preferences.
Sponsored Session
Chair: Danko Turcic, Associate Professor Of Operations, Olin Business
School, Washington University in St. Louis, St Louis, MO,
United States of America, turcic@wustl.edu
Co-Chair: Panos Kouvelis, Professor, Olin Business School, Washington
University in St. Louis, St Louis, MO, United States of America,
kouvelis@wustl.edu
334
TC50
1 - Quality at the Source or Quality at the End? Managing Supplier’s
Quality under Information Asymmetry
Mohammad Nikoofal, Católica Lisbon School of Business &
Economics, UCP, Palma de Cima, Lisbon, 1649-023, Portugal,
mohammad.nikoofal@ucp.pt
3 - Multi-item Spare Parts Inventory Planning with Selective use of
Advance Demand Information
Geert-Jan Van Houtum, Full Professor, Eindhoven University of
Technology, P.O. Box 513, Eindhoven, 5600MB, Netherlands,
g.j.v.houtum@tue.nl, Tarkan Tan, Engin Topan
In this paper, we first develop and then compare two different mechanisms for
the buyer in order to control quality improvement efforts exerted by the supplier
when the supplier has private information about his inborn reliability.
We propose a multi-item, spare parts inventory system model with a general
representation of imperfect demand information. We determine which parts
should be monitored and how much stock should be kept for each component so
that a given aggregate system availability is maintained. Our model allows excess
inventory on stock and on order to be returned to the central depot or external
supplier at a certain return cost. We also characterize the optimal ordering and
return policy.
2 - Optimal Monitoring Decisions for Asset Based Lending
Nikolaos Trichakis, Harvard Business School, Boston, MA,
United States of America, HBS, ntrichakis@hbs.edu, Dan Iancu,
Do Young Yoon
We consider a firm financing its operations by collateralizing its working assets,
e.g., inventory. To mitigate the risk due to the assets’ uncertain valuation, the
lender has a monitoring option entitling him to early repayment by liquidation.
We derive the optimal liquidation policy, showing that it can have a nonthreshold
structure. We derive bounds on the optimal monitoring time, and leverage them
to devise simple heuristics, which perform well in numerical studies.
■ TC50
50-Room 106A, CC
Operations Economics
Sponsored Session
3 - Capital Structure with Flexible Future Investments
Qi Wu, Case Western University, Cleveland, OH, United States of
America, Weatherhead School of Management, CWRU,
qxw132@case.edu, Peter Ritchken
Chair: Terry Taylor, U.C. Berkeley, Haas School of Business,
2220 Piedmont Avenue, Berkeley, CA, United States of America,
taylor@haas.berkeley.edu
We analyze the interaction between investment and financing decisions in a
dynamic contingent claims model where the firm has the ability to dynamically
control production decisions of assets in place and has growth options to invest in
that can be financed with debt and equity. The fundamental question to be
addressed is how investment timing and financing decisions are affected by the
existing capital structure and the nature of the operating flexibility inherent in
the growth options.
Co-Chair: Wenqiang Xiao, Associate Professor, New York University,
Stern School of Business, 44 West Fourth Street, 8-72, New York, NY,
10012, United States of America, wxiao@stern.nyu.edu
1 - Strategic Outscouring under Competition and
Asymmetric Information
Lusheng Shao, University of Melbourne, Melbourne, Australia,
lusheng.shao@unimelb.edu.au, Xiaole Wu, Fuqiang Zhang
4 - Make-to-Order vs. Make-to-Stock when Firms Compete, Input
Costs, and Demand are Stochastic
Danko Turcic, Associate Professor of Operations, Olin Business
School, Washington University in St. Louis, St. Louis, MO, United
States of America, turcic@wustl.edu, Guang Xiao, Panos Kouvelis
This paper studies two firms’ outsourcing strategies under competition and
asymmetric cost information. We find that without asymmetric information, the
firms will choose the supplier with smaller cost uncertainty. However, with
information asymmetry, the supplier with greater cost uncertainty may be
preferred.
This paper provides a new rationale for why firms choose long and short
production lead times that is based, in part, on non-competitive behavior in
product markets. We identify a set of conditions, which imply that some,
otherwise identical, production firms want to choose long production lead times,
while others choose short production lead times. The conditions are: (i) stochastic
production costs, (ii) price-dependent demand, and (iii) strategic inventory
withholding.
2 - Information Preferences in the Supply Chain under
Strategic Inventory
Abhishek Roy, PhD Student, McCombs School of Business,
University of Texas at Austin, 2110 Speedway Stop B6500,
Austin, TX, 78712, United States of America,
abhishek.roy@utexas.edu, Steve Gilbert, Guoming Lai
■ TC49
We investigate how the possibility of strategic inventory influences the
preferences for information sharing between supply chain partners. Among other
results, we show that the presence of strategic inventory may alter traditional
information preferences of the supply chain partners regarding the creation of a
mechanism for sharing information about the retailer’s operation with the
supplier.
49-Room 105B, CC
Multi-Echelon Inventory Modeling
Sponsored Session
3 - Product Quality in a Distribution Channel with Inventory Risk
Kinshuk Jerath, Columbia University, 521 Uris Hall, 3022
Broadway, New York, NY, 10027, United States of America,
jerath@columbia.edu, Sang Kim, Robert Swinney
Chair: Sean Willems, University of Tennessee, 453 Haslam Business
Building, Knoxville, TN, 37996, United States of America,
willems@bu.edu
1 - Velocity-based Storage in a Semi-automated Order
Fulfillment System
Stephen Graves, MIT, 77 Massachusetts Avenue, Cambridge, MA,
02139, United States of America, sgraves@mit.edu, Rong Yuan
We analyze a situation in which a product has to be designed and sold under
demand uncertainty. We consider the jointly optimal quality and inventory
decision in both a centralized channel (a single firm determines both) and a
decentralized channel (a manufacturer determines quality while a retailer
determines inventory), and discuss how demand uncertainty impacts the optimal
quality-inventory pair and how coordination of the decentralized channel may be
achieved.
Online retailers continue to invest in technology to improve the efficiency of
order fulfillment. This technology creates new operating challenges and
opportunities. We examine a semi-automated fulfillment system in which pickers
and stowers are stationary, and the inventory storage units are brought to them.
We evaluate the effectiveness of velocity-based storage and consider how to
deploy a velocity-based storage policy in light of picking, stowing and storage
decisions.
4 - Congested Platforms
Terry Taylor, U.C. Berkeley, Haas School of Business,
2220 Piedmont Avenue, Berkeley, CA, United States of America,
taylor@haas.berkeley.edu
In a platform business model, the platform firm provides a per-service wage
payment to independent agents (e.g., drivers in riding-sharing services (e.g.,
Uber), shoppers in delivery services (e.g., Instacart)) to motivate them to provide
service to customers. This paper using a queueing model to examine the impact of
congestion on the platform’s optimal price and wage.
2 - Incorporating an Operational Layer into the Guaranteed-service
Inventory Optimization Approach
Steffen Klosterhalfen, University of Richmond, 1 Gateway Road,
Richmond, VA, 23173, United States of America,
steffenklosterhalfen@googlemail.com, Daniel Dittmar
The existing guaranteed-service contributions assume bounded demand and do
not explicitly model how excess demand is handled by some type of flexibility
measure. The lack of a clear operational description leaves the material flow
representation somewhat incomplete and renders the approach controversial. We
incorporate operating flexibility in the form expediting. By doing so we can work
directly with the external (unbounded) demand and the entire material flow is
easy to trace and understand.
335
TC51
■ TC51
2 - Nonparametric Demand Predictions for New Products
Srikanth Jagabathula, NYU, 44 West Fourth Street, New York, NY,
United States of America, sjagabat@stern.nyu.edu,
Lakshminarayana Subramanian, Ashwin Venkataraman
51-Room 106B, CC
Online Retailing
Predicting demand for new products is important and challenging. Existing
parametric approaches require selection of relevant features of products and
specification of the parametric structures, both of which are challenging. We
propose a non-parametric approach combining ideas from “Learning to Rank” in
machine learning and “Choice Estimation” in operations and marketing. The
resulting methods can be used out-of-the-box and allow us to predict the impact
of changes in product features.
Sponsored Session
Chair: Dorothee Honhon, Associate Professor, University of Texas at
Dallas, 800 W. Campbell Road, Richardson, TX, 75080,
United States of America, Dorothee.Honhon@utdallas.edu
3 - A Structured Analysis of Unstructured Big Data Leveraging
Cloud Computing
Xiao Liu, Assistant Professor Of Marketing, New York University,
44 West 4th Street, New York, NY, 10012, United States of
America, xiaoliu@andrew.cmu.edu, Kannan Srinivasan,
Param Vir Singh
Co-Chair: Amy Pan, Assistant Professor, University of Florida, Dept. of
ISOM, Warrington College of Business Administr, Gainesville, FL,
32608, United States of America, amy.pan@warrington.ufl.edu
1 - Counteracting Strategic Purchase Deferrals: The Impact of Online
Retailers’ Return Policy Decisions
Tolga Aydinliyim, Baruch College, One Bernard Baruch Way,
Dept of Management Box B9-240, New York,
United States of America, Tolga.Aydinliyim@baruch.cuny.edu,
Mehmet Sekip Altug
In this study, we combine methods from cloud computing, machine learning and
text mining to illustrate how content from social media can be effectively used for
forecasting purposes. We conduct our analysis on a staggering volume of nearly
two billion Tweets. Our main findings highlight that, in contrast to basic surfacelevel measures such as volume or sentiment, the information content improve
forecasting accuracy significantly.
We study the impact of (i) forward-looking (i.e., discount-seeking) consumer
behavior and (ii) consumers’ sensitivity to clearance period stock availability on
retailers’ returns management decisions and the ensuing demand segmentation
and profit effects in both monopolistic and competitive settings.
4 - Modeling Multi-taste Consumers
Liu Liu, NYU Stern School of Business, 40 West 4th Street, Tisch
Hall, Room 825, New York, NY, 10012, United States of America,
lliu@stern.nyu.edu, Daria Dzyabura
2 - Replenishment under Uncertainty in Online Retailing
Jason Acimovic, Penn State University, 462 Business Building,
University Park, PA, 16802, United States of America,
jaa26@smeal.psu.edu, Stephen Graves
In many product categories where recommendation systems are used, a single
consumer may have multiple different tastes. We propose a framework for
modeling choice behavior of such a multi-taste consumer and an iterative
algorithm for estimation. We test it in numerical studies and an empirical
application (Allrecipes.com). Our results show that it has superior out-of-sample
predictive performance than single-taste models and is able to accurately recover
parameters in simulation studies.
Online retailers often may serve most customers from any warehouse location.
Simple order-up-to policies are easy to implement; however, they may perform
suboptimally leading to high shipping costs. We propose a replenishment heuristic
based on bringing inventory up to a target level on the day inventory arrives. We
calculate robust target levels, taking into account demand uncertainty, shipping
costs, and estimated stockout costs. We show how this policy performs on realistic
data.
3 - Optimal Spending for a Search Funnel
Shengqi Ye, The University of Texas at Dallas, 800 West Campbell
Road, Richardson, TX, 75080, United States of America,
sxy143530@utdallas.edu, Goker Aydin, Shanshan Hu
■ TC53
53-Room 107B, CC
Grab Bag of Behavioral Papers
Sponsored search marketing has been a major advertising channel for online
retailers. Recent observation indicates that not all customers finalize their
purchase decision after their first search query. Instead, customers might take a
path of keywords and clicks - a search funnel - to complete a conversion. Noting
this behavior, we investigate a retailer’s optimal advertising budget allocation
across keywords in the search funnel.
Sponsored Session
Chair: Kenneth Schultz, Associate Professor, AFIT,
2950 Hobson Way, WPAFB, OH, 45433, United States of America,
Kenneth.Schultz@afit.edu
1 - The Influence of Education and Experience Upon Contextual and
Task Performance in Warehouse Operations
Allen Miller, Student, Air Force Institute of Technology,
2950 Hobson Way, WPAFB, OH, 45433, United States of America,
Allen.miller@afit.edu
4 - Omnichannel Inventory Management with
Buy-Online-and-Pickup-in-Store
Fei Gao, The Wharton School, University of Pennsylvania, 3730
Walnut Street, 500 Jon M. Huntsman Hall, Philadelphia, PA,
United States of America, feigao@wharton.upenn.edu,
Xuanming Su
Many retailers offer customers the option to buy online and pick up orders in
store. We study the impact of this omnichannel strategy on store operations and
offer recommendations to retailers.
We believe worker-performance may be affected by the individual’s knowledge of
why and where they fit into a larger system, defined as mission clarity. We
conduct a controlled experiment to discern how education, experiences and
subject characteristics impact mission clarity and subsequently contextual and
task performance in a pick-and-pack operation.
■ TC52
2 - Personal Bias and Contract Setting
Julie Niederhoff, Syracuse University, Syracuse, NY
United States of America, jniederh@syr.edu
52-Room 107A, CC
Machine Learning Applications in Marketing
The efficacy and necessity of coordinating contracts has a strong analytical
support, but experimental work shows that decision makers do not set the
contracts as theory prescribes. Alternative objective functions of risk and fairness
are explored at the individual level to understand when and for whom the
contracts are most and least effective and necessary.
Sponsored Session
Chair: Daria Dzyabura, Assistant Professor of Marketing, NYU Stern
School of Business, 40 West 4th Street, Tisch 805, New York, NY,
10012, United States of America, ddzyabur@stern.nyu.edu
1 - Big Data Pricing
Eric Schwartz, University of Michigan, Ann Arbor, Michigan,
United States of America, ericmsch@umich.edu, Kanishka Misra
3 - The Influence of Emergency Medical Services Load on
Paramedics On-scene Clinical Decisions
Mohammad Delasay, Post-doctoral Fellow, Tepper School of
Business, Carnegie Mellon University, 5000 Forbes Avenue,
Pittsburgh, PA, United States of America,
delasays@andrew.cmu.edu, Kenneth Schultz, Armann Ingolfsson
We study how a firm should maximize revenue by dynamically setting its price
over time for a new product. We solve this optimal experimentation problem as a
multi-armed bandit problem combined with economic theory. The approach adds
to dynamic pricing in marketing and econometrics using non-parametric
identification of demand by using reinforcement learning. In particular, we derive
a pricing algorithm based on upper confidence bound and illustrate its theoretical
and empirical properties.
We investigate the effect of emergency medical system load on paramedics’
medical decisions. We hypothesize that paramedics’ decisions about on-scene time
and transporting a patient to hospital are influenced by the emergency system
load. We test our hypotheses by analyzing a data set of emergency responses in
Calgary, Canada.
336
TC56
4 - The Bright Side of Managerial Overconfidence
Juan Li, Assistant Professor, Nanjing University, No. 5 Ping Cang
Xiang, Nanjing, China, juanli@nju.edu.cn, Baojun Jiang,
Fuqiang Zhang
4 - Opportunism in Manufacturing Outsourcing
Keith Skowronski, The Ohio State University, Fisher 251A, 2100
Neil Avenue, Columbus, OH, 43210, United States of America,
skowronski.2@osu.edu, W. C. Benton
Managers are often overconfident about the accuracy of their demand forecast.
This paper shows that the firm may actually benefit from such overconfidence
bias whether or not its competitor has such bias. Further, such bias can lead to a
win-win situation for both competing firms.
Using dyadic buyer-supplier data, we empirically examine two types of supplier
opportunism, poaching and shirking, in manufacturing outsourcing relationships.
In this multi-country study, the legal environment of the supplier’s location is
hypothesized to moderate the relationships between exchange hazards, relational
governance mechanisms and the different forms of opportunism.
5 - Suppliers as Liquidity Providers
Panos Markou, IE Business School, Calle Maria de Moina 12
Bajo, Madrid, 28006, Spain, pmarkou.phd2016@student.ie.edu,
Daniel Corsten
■ TC54
54-Room 108A, CC
Discrete Optimization Models for Homeland Security
and Disaster Management
Using a novel dataset comprising the top 10 suppliers of more than 80,000 public
and private companies, we examine the value of having a supplier that is not
financial constrained. For financially constrained customers, holding cash is costly.
However, we show that having even one financially unconstrained supplier
allows these customers to “outsource” some of their cash holdings. Financial
standing is an important consideration when choosing suppliers.
Cluster: Tutorials
Invited Session
Chair: Laura Mclay, Associate Professor, University of Wisconsin,
1513 University Ave, ISYE Department, Madison, WI, 53706,
United States of America, lmclay@wisc.edu
1 - Discrete Optimization Models for Homeland Security and
Disaster Management
Laura Mclay, Associate Professor, University of Wisconsin,
1513 University Ave, ISYE Department, Madison,WI, 53706,
■ TC56
56-Room 109A, CC
Commercialization of New Technologies
Invited Session
Preparing for and responding to disasters, including acts of terrorism, is an
important issue of national and international concern. Recent disasters
underscore the need to manage disasters to minimize their impact on critical
infrastructure and human suffering. In this tutorial, we survey the operations
research literature that develops, analyzes and applies discrete optimization
models to effectively mitigate, prepare for, respond to and recover from a wide
variety of disasters.
Chair: Karthik Ramachandran, Georgia Institute of Technology, 800
West Peachtree NW, Atlanta, GA, 30308, United States of America,
Karthik.Ramachandran@scheller.gatech.edu
Co-Chair: Sreekumar Bhaskaran, SMU, Dallas, TX,
United States of America, sbhaskar@cox.smu.edu
1 - Product Line Design for Strategic Customers
Saurabh Bansal, Assistant Professor, Penn State Univrsity, 405
Business Building, University Park, PA, 16802, United States of
America, sub32@psu.edu, Karthik Ramachandran
■ TC55
55-Room 108B, CC
We report results for optimal product line design when customers are strategic
about uncertain quality of products. Our analysis explains evolution of product
lines observed in practice.
Outsourcing I
Contributed Session
2 - Licensing Contracts: Control Rights and Options
Niyazi Taneri, SUTD, 8 Somapah Rd, Singapore, Singapore,
niyazitaneri@sutd.edu.sg, Pascale Crama, Bert De Reyck
Chair: Ting Luo, University of Texas at Dallas, 800 W Campbell Rd,
Richardson, TX, 75080, United States of America, ting.luo@utdallas.edu
1 - Outsourcing Supplier Selection: Quality-driven Demand and
Taguchi Loss Function
Yanni Ping, Drexel University, 3220 Market Street, Gerri C.
LeBow Hall 730, Philadelphia, PA, 19104, United States of
America, yp86@drexel.edu, Seung-lae Kim, Min Wang
Research and development (R&D) collaborations, though common in high-tech
industries, are challenging to manage due to technical and market risks as well as
incentive problems. We investigate the impact of control rights, options, payment
terms and timing decisions on R&D collaborations between an innovator and a
marketer. We provide recommendations on the optimal contract structure and
timing based on the R&D project characteristics.
Facing limited capacity, a manufacturer would often rely on external suppliers.
How to select suppliers to work with becomes a strategic decision particularly
when demand for the final product is quality driven. In this talk, we adopt a
Taguchi loss function as a supplier’s quality measurement and present a dynamic
programming model to explore how supplier quality affects manufacturer’s
outsourcing strategy. We propose simple and efficient algorithms for supplier
selection in a dynamic setting.
3 - Does Equity Crowdfunding Improve Entrepreneurial
Firm Performance?
Susanna Khavul, UTA/London School of Economics, London,
United Kingdom, s.khavul@lse.ac.uk, Saul Estrin
As a fast moving financial innovation, equity crowdfunding may relax resource
constraints for new ventures. Using four years of proprietary data, we model how
information provision, generation, and exchange affects the supply of funds and
likelihood of pitch funding. We evaluate this against the survival and performance
of the firms that sought funding.
2 - Fixed Entry Cost Effect on Contract Length and Renewals in a
Maintenance Service Contract Systems
Rodrigo Ulloa, Pontificia Universidad Católica de Chile, Av.
Vicuña Mackenna 4860, Santiago, Chile, rsulloa@uc.cl,
Alejandro Mac Cawley, Rodrigo Pascual, Gabriel Santelices
We analyze how the inclusion of a fixed entry cost will affect the decision making
of a maintenance contract, using a model that evaluates the contract value for the
vendor according to the contract duration and its renewals. The analysis considers
different scenarios that show the existence of a relationship between the length of
the contract and the amount of renovations from which the contract is valuable
for the vendor.
3 - Long-term Outsourcing under Stochastic Learning and
Information Asymmetry
Ting Luo, University of Texas at Dallas, 800 W Campbell Rd,
Richardson, TX, 75080, United States of America,
ting.luo@utdallas.edu
Suppliers can reduce their cost through learning by doing, however their learning
abilities and outcomes are kept as private information. When buyers design the
procurement contract, they must consider the above effects. We study the
interplay of stochastic learning and information asymmetry. We show that the
stochastic learning has a profound impact on the optimal contract.
337
TC57
■ TC57
drought risk assessment model is constructed from the thermoelectric power
sector’s operational perspective.
57-Room 109B, CC
2 - Portfolio Approach for Optimal Rooftop Solar Arrays Selection for
Distributed Generation
Olufemi Omitaomu, Senior Research Scientist, Oak Ridge
National Laboratory, 1 Bethel Valley Road, MS-6017, Oak Ridge,
TN, 37831, United States of America, omitaomuoa@ornl.gov,
Xueping Li
Long-Term Electric Power System Planning Models
Sponsor: ENRE – Energy I – Electricity
Sponsored Session
Chair: Ramteen Sioshansi, Associate Professor, The Ohio State
University, Integrated Systems Engineering, 1971 Neil Avenue,
Columbus, OH, 43210, United States of America, sioshansi.1@osu.edu
1 - Optimizing Storage Operations in Longer-term Power
System Models
Sonja Wogrin, Comillas Pontifical University, Calle Alberto
Aguilera 23, Madrid, Spain, sonja.wogrin@comillas.edu,
David Galbally, Javier Reneses
We present a portfolio selection approach that consider thousands of buildings
with different solar energy potential and that are being considered for utility-scale
distributed power generation. Our approach uses Markowitz mean-variance
portfolio selection model to select suitable rooftops by identifying a combination
of buildings that will maximize solar energy outputs and minimize system
variability. Our approach is implemented using some real data-sets.
In a rapidly changing power system the proper characterization of storage
behavior becomes an increasingly important issue. We propose a new formulation
to capture storage behavior in medium- and long-term power system models that
use a load duration curve. In such models the chronology among individual hours
is lost; our approach addresses related shortcomings and is able to accurately
replicate hourly evolution of storage levels while keeping computational time
tractable.
■ TC59
59-Room 110B, CC
Impacts of Climate Change
Sponsored Session
2 - Multi-stage Investment in Renewable Energies via Linear
Decision Rules
Maria Ruth Dominguez Martin, Universidad de Castilla La Mancha, Av. Carlos III, s/n, Toledo, Spain,
Ruth.Dominguez@uclm.es, Miguel Carrion, Antonio Conejo
Chair: Chris Lauer, Oregon State, Portland, OR,
United States of America, cjlauer@gmail.com
1 - Planning Forest Harvesting under Climate Change:
A Stochastic Optimization Model
Jordi Garcia, Researcher, Instituto Superior de Agronomia,
Universidade de Lisboa, Portugal, Tapada da Ajuda., Lisbon,
Portugal, jordigarcia@isa.ulisboa.pt, Andres Weintraub,
Cristobal Pais, Joanna Bachmatiuk
Investment in generating capacity involves high uncertainty. In the real world,
these decisions are usually made in several stages. We propose a multi-stage
investment model to transform a thermal-based power system into a renewabledominated one. We consider the uncertainty of the demand growth and the
investment costs, as well as the variability of the renewable power production
throughout the year.
In this paper we consider a medium term forest planning problem in the presence
of uncertainty due to climate change. For each time period the forest planner
must decide which areas to cut in order to maximize expected net profit. A
multistage stochastic model using 32 climate scenarios was developed and solved
to determine optimal harvesting decisions under uncertainty. The stochastic
solutions were compared to the solution of a deterministic model where an
average climate scenario was used.
3 - Impact of Unit Commitment Constraints on Generation Expansion
Decisions under Wind Uncertainty
Jalal Kazempour, Technical University of Denmark, Department
of Electrical Engineering, Kgs. Lyngby, 2800, Denmark,
seykaz@elektro.dtu.dk, Amin Nasri, Antonio Conejo
2 - Will Climate Change Induced Effects Cause Harm to Value Chains
of the Bio-based Industry?
Peter Rauch, Pd, BOKU, Feistmantelstrasse 4, Wien, 1180,
Austria, peter.rauch@boku.ac.at
Most of available generation expansion decision-making tools in the literature
neglect a number of unit commitment (UC) constraints, e.g., on/off commitment
status, ramping limits and minimum production level of thermal units. This study
analyzes the impact of those constraints on generation expansion decisions in
power systems with significant wind power integration. To this end, a two-stage
stochastic programming problem is proposed and solved using a decomposition
technique.
Uncertainty increasingly affects ecosystems and storms and bark beetle
infestations are the main causes of forest damage. Risks and their impacts on
value chains of the bio-based industry are evaluated by a System Dynamics model
of the Austrian wood supply including a stochastic simulation of risk agents.
Results provide insights on probabilistic future wood supply security for sawlogs,
pulpwood resp. energy wood and reveal a contra-intuitive system effect for the
climate change scenario.
4 - Stochastic Generation and Transmission Investment
Planning Model
Yixian Liu, The Ohio State University, 210 Baker Systems Bldg.,
1971 Neil Ave., Columbus, OH, 43210, United States of America,
liu.2441@osu.edu, Ramteen Sioshansi
3 - Multiobjective Optimization to Study the Impact of Climate
Change on the Joint Provision of Ecosystem Services
Nick Kullman, Masters Student, University of Washington,
360 Bloedel Hall, Seattle, WA, United States of America,
nick.kullman@gmail.com, Sventlana Kushch, Sandor Toth
Increasing electricity demand makes it necessary to expand the capacity of the
current electrical grid. We propose a multi-stage stochastic linear programming
model to seek an optimal investment plan in generators, storage devices and
transmission lines in a long time horizon. Multiple uncertainties are considered
involving both investment and dispatch decisions. The model is large in scale and
may take excessive computational time to be solved. Decomposition methods will
be applied to solve it.
Climate change has been shown to alter the provision of forest ecosystem services
such as carbon sequestration and wildfire mitigation. Less understood is how
climate change will alter the tradeoffs among ecosystem services acquired
simultaneously. We present a scenario-based multi-objective mathematical
programming method to study these changes on the joint provision of ecosystem
services in the Deschutes National Forest.
■ TC58
4 - Incorporating Acclimation and Feedback into Reserve Selection
During Climate Change
Austin Phillips, University of Washington, Seattle, WA, United
States of America, austinjphillips90@gmail.com, Sandor Toth,
Robert Haight
58-Room 110A, CC
Electricity and System Resilience
Sponsored Session
Climate change threatens many species, and conservation in such a dynamic
setting is challenging. We developed a mixed integer reserve selection model that
pairs population dynamics and sequential selection in a nonlinear feedback loop.
The model accounts for species’ ability to acclimate, as well as disperse, to track
suitable conditions. We explore optimal management strategies to facilitate
species’ survival as they disperse and acclimate in response to warming.
Chair: Valerie Thomas, Professor, Georgia Institute of Technology,
755 Ferst Drive, NW, Atlanta, GA, United States of America,
valerie.thomas@isye.gatech.edu
1 - Stochastic Model for Assessment of Thermoelectric Power
Generation Drought Risk under Climate Change
Royce Francis, George Washington University, 800 22nd St. NW
B1850, Washington, DC, 21212, United States of America,
seed@gwu.edu, Behailu Bekera
The objective of this article is to propose a stochastic method for analyzing
drought risk to the thermoelectric power generation infrastructure sector due to
its heavy reliance on freshwater availability. In particular, this article proposes a
thermoelectric drought characterization framework from which a stochastic
338
■ TC60
■ TC61
60-Room 111A, CC
61-Room 111B, CC
Disruption Management
Sustainable and Responsible Supply Chain
Management
Contributed Session
TC62
Sponsored Session
Chair: Min Ouyang, Associate Professor, Huazhong University of
Science and Technology, Room W308 in S1 Building, 1037 Luoyu
Road, Wuan, 430074, China, mouyang618@gmail.com
1 - Transportation Network Protection: A Model with Variable Flow
Demand
Stefano Starita, PhD Researcher, Kent Business School, University
of Kent, University of Kent, Canterbury, CT2 7PE, United
Kingdom, s.starita@kent.ac.uk, Dr. Maria Paola Scaparra
Chair: Jose Cruz, Associate Professor, University of Connecticut,
2100 Hilllside Road, Storrs, CT, 06269, United States of America,
Jose.Cruz@business.uconn.edu
1 - Corporate Social Responsibility and Supply Chain Profitability
Zugang Liu, Associate Professor, Penn State Hazleton, 76
University Dr, Hazleton, PA, United States of America,
zxl23@psu.edu, Trisha Anderson, Jose Cruz
Protecting transportation infrastructure is critical to avoid life and economic
losses. We model a fortification problem on an all-pairs, flow-based network. To
model system users’ behavior, the traffic demand is assumed to be dependent on
the length of the shortest path available. We present an efficient heuristic solution
approach and a case study on the London tube.
We find that environmental and social responsible activities have different
impacts on different stages of supply chains. For manufacturers, positive social
activities and negative environmental activities increase the return on assets; for
wholesalers, neither social nor environmental activities has significant impact; for
retailers, negative environmental activities negatively affect the return on assets.
2 - Comparison of Supply Chain Recovery Policies After a
Major Disruption
Joanna Marszewska, Assistant Professor, Jagiellonian University,
Department of Japanology and Sinology, Krakow, 31120, Poland,
rokimi@op.pl, Tadeusz Sawik
2 - The Amazon Tax and E-tailer Supply Chains
Trisha Anderson, Associate Professor, Texas Wesleyan University,
1201 Wesleyan Street, Fort Worth, TX, United States of America,
trdanderson@txwes.edu, Kevin Mcgarry
Different recovery policies of a supply chain after major disruption caused by
natural disasters are presented. The Japan’s competiveness-robustness dilemma is
discussed against a resilient supply chain design strategy. Single, dual or multiple
sourcing, improved suppliers visibility, protection of suppliers against natural
disasters and prepositioning of emergency inventory of product-specific parts
along a supply chain are considered and their impact on the recovery process is
analyzed.
We study two hypothesis to address a key legal question that e-tailers consider
when opening up distribution centers: whether they should operate under the
assumption that collecting state sales tax for all online transactions is inevitable or
continue to strategically position themselves to minimize the tax burden where
possible, even if it compromises supply chain strategic positioning. We also study
the environmental implications of these e-tailer supply chain decisions.
3 - Cargo Prioritization and Terminal Allocation in Case of Inland
Waterway Disruption
Liliana Delgado Hidalgo, Graduate Student, University of
Arkansas, 4207 Bell Engineering Center, Fayetteville, AR, 72701,
United States of America, ld002@uark.edu, Heather Nachtmann
3 - Social Responsibility Investments: Financial Networks,
Transaction Cost, and Risk Effects
Jose Cruz, Associate Professor, University of Connecticut, 2
100 Hilllside Road, Storrs, CT, 06269, United States of America,
Jose.Cruz@business.uconn.edu
We propose a solution approach to reroute barges in case of an Inland waterway
disruption. The first part of the solution uses an Analytic Hierarchical Process
(AHP) to assign priority index to the barges. We formulate a Integer Linear
Problem to assign the barges to the terminals where the cargo is offloaded to be
transported by a different transportation mode. The AHP results are used to
schedule the barges assigned to a terminal. A case example is presented to
illustrate our results.
This paper develops a network equilibrium model in conjunction with capital
asset pricing model (CAPM) and the net present value (NPV) to determine the
optimal portfolio, prices, profits, and equity values of financial network firms
under financial risks and economic uncertainty. We investigate how social
responsible financial investment decisions affect the values of interconnected
financial firms from a network perspective.
4 - Green Building Decision-making using an Exploration and
Exploitation Approach
John Dickson, Symphony Teleca Analytics, 5360, Legacy Drive,
Plano, TX, United States of America, john.dickson@mavs.uta.edu,
Jay Rosenberger, Victoria Chen
4 - Resilient Design in Agribusiness Supply Chain under
Supply Disruptions
Golnar Behzadi, PhD Student, University of Auckland, Level 2,
Room 439-215,70 Symonds St, Auckland, 1010, New Zealand,
gbeh681@aucklanduni.ac.nz, Abraham Zhang, Tava Olsen,
Michael O’sullivan
The experiments or simulations conducted by computers can be a tedious task,
which require substantial computational time. This research focuses on
developing a surrogate based optimization, in which we iteratively build a
surrogate model, using few points and then optimize the model by adding more
points until the best solution is found. A single story residential green building
based in California is used as a case study.
Agribusiness supply chains have limited lifecycle of products, seasonality of
supply and demand, long lead time for production and delivery, and supply that is
affected by climatic variability, which makes them especially vulnerable to supply
disruptions. A special approach to risk management is required and here we
consider resilience. Resilience incorporates concepts from vulnerability and risk
management to address the recovery of a system from disruptions (rare highimpact risks).
■ TC62
5 - Decision Support for Critical Infrastructure
Resilience Enhancement
Min Ouyang, Associate Professor, Huazhong University of Science
and Technology, Room W308 in S1 Building, 1037 Luoyu Road,
Wuan, 430074, China, mouyang618@gmail.com
62-Room 112A, CC
Optimization in Bio-energy
Cluster: Energy Systems: Design, Operation, Reliability and
Maintenance
Invited Session
It develops a framework of resilience decision support system RDSS for critical
infrastructures. This RDSS includes seven modules: Data Input, Property Statistics,
Scenario Generation, Vulnerability Analysis, Restoration Simulation, Resilience
Assessment and Strategy Exploration, which together allow for statistically and
visually exploration of critical infrastructure system resilience under point and
period disruption scenarios and facilitates effectiveness analysis of resilience
strategies.
Chair: Mohammad Marufuzzaman, Mississippi State University,
Industrial & Systems Engineering, Starkville, MS, 39762,
United States of America, mm2006@msstate.edu
1 - Designing a Dynamic Multimodal Transportation Network under
Biomass Supply Uncertainty
Sushil Poudel, Mississippi State University, Starkville, MS,
United States of America, srp224@msstate.edu, Mohammad
Marufuzzaman, Linkan Bian, Hugh Medal
This study presents a two-stage stochastic programming model that assigns multimodal facilities dynamically to design a biomass supply chain network under
feedstock supply uncertainty. We develop algorithms combining sample average
algorithm, progressive hedging algorithm, and rolling horizon algorithm to solve
this challenging NP-hard problem.
339
TC63
2 - Designing a Reliable Bio-fuel Supply Chain Network Considering
Link Failure Probabilities
Linkan Bian, Assistant Professor, Mississippi State University,
260 McCain Building, Starkville, MS, 39762, United States of
America, bian@ise.msstate.edu, Sushil Poudel,
Mohammad Marufuzzaman
5 - Progressive Modeling: Towards a New Complex Systems
Optimization Paradigm
Mohamed Ismail, Assistant Professor, University of Regina,
3737 Wascana Parkway, Regina, SK, S4S0A2, Canada,
mohamed.ismail@uregina.ca
Progressive Modeling (PM) is a multidisciplinary forward-looking modeling
approach that finds pragmatic solutions for many complex and large-scale
industrial problems. Many related applications will presented to demonstrate the
principles and the techniques adopted in this paradigm. The new modeling
paradigm is expected to have many engineering applications and influence many
disciplines such as systems optimization, Operations management, and system of
systems engineering.
This study presents a pre-disaster planning model that seeks to strengthen the
multi-modal facilities links for a bio-fuel supply chain system under limited
budget availability. The failure probability of the links are estimated using a
spatial-statistical model. We developed a combinatorial Benders decomposition
algorithm to solve this challenging NP-hard problem.
3 - Managing Congestion in a Multi-modal Facility Location Problem
under Uncertainty
Mohammad Marufuzzaman, Mississippi State University,
Industrial & Systems Engineering, Starkville, MS, 39762,
United States of America, mm2006@msstate.edu
■ TC64
64-Room 113A, CC
This paper presents a mathematical model that studies the impacts of the
congestion effect in a multi-modal facility location design problem under
feedstock supply uncertainty. The model is solved using a hybrid algorithm that
integrates constraint generation, sample average approximation, progressive
hedging and rolling horizon algorithm.
Panel Discussion: Analytics and Decision Analysis
Sponsored Session
Chair: Jeffrey Keisler, University of Massachusetts Boston,
1 - Analytics and Decision Analysis
Moderator:Jeffrey Keisler, University of Massachusetts Boston,
100 Morrissey Boulevard, Boston, MA, 02125, United States of
America, Jeff.Keisler@umb.edu, Panelists: Jeff Vales,
Casey Lichtendahl, John Turner, Don Kleinmuntz, Max Henrion
■ TC63
63-Room 112B, CC
Operations Management I
Contributed Session
Chair: Mohammed Darwish, Associate Professor, Kuwait University,
Industrial and Management Systems Eng., P.O. Box 5969, Safat, 13060,
Kuwait, m.darwish@ku.edu.kw
1 - Probabilistic Estimation of the Inventory Shortage Cost
Feng Xu, Georgia Southwestern State University, 800 GSW State
University Drive, School of Business Administration, Americus,
GA, 31709, United States of America, feng.xu@gsw.edu
Huge increases in data availability and computing power have transformed
quantitative fields and led to a proliferation of tools for analytics. Panelists will
discuss how can DA strengthen analytics broadly defined, and how can analytics
strengthen DA.
■ TC65
Due to the difficulty in calculating the loss of goodwill, in estimating the shortage
cost practitioners and researchers often assume a fixed penalty cost or switch to
assigning a specific customer service level. This paper proposes probabilistic
measurements of the shortage cost, based on mathematical relationship between
the cost and the shortage amount. The derived closed-form estimates of the
expected shortage cost can then be applied to determining the optimal inventory
control policy.
65-Room 113B, CC
Joint Session DAS/MAS:Game Theory, Decision
Analysis, and Homeland Security, Part B
Sponsored Session
2 - Optimal Staffing with Endogenous Goals
Buket Avci, Singapore Management University, 50 Stamford
Road, Singapore, 248196, Singapore, buketavci@smu.edu.sg
Chair: Jun Zhuang, University at Buffalo, SUNY, 317 Bell Hall, Buffalo,
NY, 14221, United States of America, jzhuang@buffalo.edu
1 - Modeling A Multi-target Attacker-defender Resource Allocation
Game Considering Risk Preferences
Jing Zhang, University at Buffalo, SUNY, 338 Bell Hall, Buffalo,
NY, 14221, United States of America, jzhang42@buffalo.edu,
Jun Zhuang, Victor Richmond Jose
We investigate the optimal staffing level decision of a firm, when employee
performance is indirectly affected by staffing levels through workload. In the spirit
of Prospect Theory, we posit that goals act as reference points, and there is an
asymmetry between under and over-performance relative to a goal. We solve the
corresponding principal-agent model in a queueing context and characterize
conditions when endogenous goals are relevant for staffing decisions.
Although evidence has been found that people often demonstrate risk preference
when faced with risky decisions, the literature mostly assumed that adversaries
are risk-neutral. This paper models a sequential attacker-defender game where
the defender allocates defensive resources to multiple targets while considering
the risk preferences of both the defender and attacker. We study the cases when
the attacker could be either non-strategic, or strategic.
3 - Quality Management Theory Development via Meta-analysis
Xianghui Peng, University of North Texas, 1307 West Highland
Street, College of Business, Denton, TX, 76201, United States of
America, xianghui.peng@unt.edu, Victor Prybutok, Robert Pavur
A meta-analysis is conducted on the empirical studies in quality management
(QM). The results allow evaluation of the relationship strength among QM
practices, performance, and content factors. The longitudinal evaluation in this
study investigates how relationships and content factors in the post-2005 period
compare with the pre-2005 period.
2 - Game Theoretic Analysis of Secret and Reliable Communication
Melike Baykal-görsoy, Rutgers, The State University of New
Jersey, 96 Frelinghuysen Road, CoRE Building, Room 201,
Piscataway, NJ, 08854, United States of America,
gursoy@rci.rutgers.edu
4 - Determination of the Maximum Worth of Auctioned Lots using
Acceptance Sampling Method
Mohammed Darwish, Associate Professor, Kuwait University,
Industrial and Management Systems Eng., P.O. Box 5969, Safat,
13060, Kuwait, m.darwish@ku.edu.kw, Fawaz Abdulmalek
Secret and reliable communication presents a challenge involving a double
dilemma for a user and an adversary. To get insight into this problem, we present
two simple stochastic games. Explicit solutions are found. In addition, we show
that under some conditions, incorporating in the transmission protocol a time slot
dealing just with the detection of malicious threats can improve the secrecy and
reliability of the communication without extra transmission delay.
In recent years, auction becomes an important method of buying and selling
different items around the world. The most common type of auctions that is
found in practice is the English Auction where a bidder inspects the auctioned lot
by taking a sample and based on the number of defective items found in the
sample, he or she takes a critical decision regarding the maximum worth of the
auctioned lot. We show how the maximum worth of an auctioned lot can be
determined using acceptance sampling.
3 - Optimal Cost-sharing in General Resource Selection Games
Konstantinos Kollias, Stanford University, 474 Gates Building,
353 Serra Mall, Stanford, CA, 94305, United States of America,
kkollias@stanford.edu, Tim Roughgarden, Vasilis Gkatzelis
Resource selection games provide a model for a diverse collection of applications
where a set of resources is matched to a set of demands. In reality, demands are
often selfish and congestion on the resources results in negative externalities for
their users. We consider a policy maker that can set a priori rules to minimize the
inefficiencies induced by selfish behavior and we characterize the control methods
that minimize the worst-case inefficiency of equilibria.
340
4 - Strategic Security Screening Queue with Abandonments
Ali Pala, PhD Student, University at Buffalo, 441 Delta Rd, Apt. 2,
Amherst, NY, 14226, United States of America,
alipala@buffalo.edu, Jun Zhuang
■ TC67
Risk of threats and congestion are the major problems in security screening
procedures. Strict security screening policies help detect or deter the adversary
applicants, but also lead to congestion which may discourage good applicants
from applying or cause unnecessary abandonment from the screening queue. This
research focuses on a novel strategic queueing system and proposes a data
supported game-theoretical model to study this problem.
Sponsored Session
TC68
67-Room 201A, CC
Consolidation in Transport
Chair: Wentao Zhang, University of Southern California,
Los Angeles, CA, United States of America, United States of America,
wentao@usc.edu
1 - Frequency-Location Clustering for Efficient Inbound Routes to
Consolidation Centers
Zhijie Dong, Cornell University, United States of America,
zd57@cornell.edu, Mark A. Turnquist
5 - When Security Games Go Green: Designing Defender Strategies
to Prevent Poaching and Illegal Fishing
Fei Fang, University of Southern California, 941 Bloom Walk,
SAL 300, Los Angeles, CA, United States of America,
feifang@usc.edu, Milind Tambe, Peter Stone
Building on the successful applications of Stackelberg Security Games (SSGs) to
protect infrastructure, researchers have begun focusing on applying game theory
to green security domains such as protection of endangered animals and fish
stocks. We introduce Green Security Games (GSGs), a novel game model for
green security domains and provide algorithms to plan effective sequential
defender strategies and to learn adversary models that further improves defender
performance.
An optimization model addresses joint decisions of frequency of pickup from
individual suppliers and grouping suppliers into collection routes by clustering in
both time and space. The objective is to minimize total logistics (transportation
plus inventory) cost. The optimization problem is equivalent to a single-source
fixed charge facility location problem, and near-optimal solutions are found using
a very efficient heuristic algorithm. Results of numerical experiments show the
effectiveness of both the model formulation and the heuristic solution method. A
case study demonstrates that substantial total cost savings can be achieved in
realistic applications using the combined frequency-location clustering method.
■ TC66
2 - A Lagrangian-based Strategy to Consolidate Freight of
Perishable Products
Christine Nguyen, Northern Illinois University, DeKalb, IL, United
States of America, United States of America, cnguyen@niu.edu,
Alejandro Toriello, Maged Dessouky
66-Room 113C, CC
Managing Airport Arrival Flows
Sponsored Session
Our research focuses on a supply chain of suppliers with low demand for
perishable products, where consolidating their product would take advantage of
better shipping FTL rates versus LTL or courier rates. We consider a Lagrangian
Relaxation formulation that includes a capacity constraint for a shared
consolidation center. We develop an LR-based heuristic that aims to balance the
consolidated economical shipping cost and the inventory cost at the consolidation
center.
Chair: John-Paul Clarke, Georgia Tech, 270 Ferst Drive, N.W.,
Atlanta, GA, United States of America, johnpaul@gatech.edu
1 - Combining Control by CTA and Enroute Speed Adjustment to
Improve Ground Delay Program Performance
James Jones, University of Maryland, 3117 A.V Williams, College
Park, MD, 20742, United States of America, jonesjc1@umd.edu,
Michael Ball, David Lovell
3 - Cost Sharing Mechanism Design for Supply Chain Consolidation
and Cooperation in Agriculture Industry
Wentao Zhang, University of Southern California, Los Angeles,
CA, United States of America, United States of America,
wentao@usc.edu, Nelson Uhan, Alejandro Toriello,
Maged Dessouky
Over the past several years there have been proposals and discussions regarding a
move from the use of controlled times of departure (CTDs) to controlled times of
arrival (CTAs) for ground delay programs (GDPs) in the U.S. In this talk we show
that, by combining control by CTA with the judicious use of en route speed
control, significant improvements to GDP performance can be achieved.
We design cost sharing mechanisms for a consolidation center where suppliers
who need to ship products to a common destination can consolidate their
shipments and save transportation costs. Using the Moulin mechanism
framework, we propose cost sharing mechanisms that are group strategyproof and
budget-balanced. By studying the efficiency of these mechanisms empirically and
analytically, we show that the outcome of these mechanisms often closely
resembles an optimal solution of a central planner.
2 - Robust Airport Gate Planning – First Order Stability Concept
Bruno Santos, Assistant Professor, TU Delft, Faculty of Aerospace
Engineering, Delft, Netherlands, B.F.Santos@tudelft.nl,
Dennis Buitendijk, Joris De Kaey, John-Paul Clarke
We present a novel approach to the airport gate assignment problem entitled
“First Order Stability”(FOS). The FOS has the goal of increasing gate plans
robustness and uses two key concepts to achieve this: it postpones the gate
scheduling to a moment when uncertainty is reduced significantly; and it
stabilizes the order of flights, minimizing the risk on disturbances. A real case
study application showed that FOS provides more stable solutions that can make
equal or higher usage of the capacity.
4 - Temporal Shipment Consolidation under Stochastic
Dynamic Demand
Sila Cetinkaya, SMU, EMIS and ITOM Departments, Dallas, TX,
United States of America, sila@smu.edu, Liqing Zhang
We consider stochastic dynamic shipment consolidation problems with general
demands and characterize the structural properties of optimal shipment release
policies under general cost structures with scale economies and quantity
discounts.
3 - Heuristic Gate Assignment Model for Airports with Multiple
Parallel Concourses
Parth Shah, Graduate Research Assistant, Georgia Tech, 401 17th
Street, Apt. 5205, Atlanta, GA, 30363, United States of America,
parth.shah1053@gmail.com, John-Paul Clarke
■ TC68
Ramp operation model of Atlanta International airport is simulated to understand
the characteristic of aircraft movement on ramp. A new heuristic approach is
adopted in which aircraft are assigned gates based on their direction of ramp entry
and exit points. The model is calibrated using ASPM traffic data. The results show
that the proposed new method achieves 23% reduction in total ground delay by
significantly reducing the gate wait, taxi blocking and pushback blocking time.
68-Room 201B, CC
TSL Prize Session
Sponsored Session
Chair: Barrett Thomas, Associate Professor, The University of Iowa,
W272 Pappajohn Business Building, Iowa City, IA, 52242,
United States of America, barrett-thomas@uiowa.edu
1 - TSL Prize Winners
Barrett Thomas, Associate Professor, The University of Iowa,
W272 Pappajohn Business Building, Iowa City, IA, 52242,
United States of America, barrett-thomas@uiowa.edu
The TSL 2015 Prize Session finalists will present their award-winning work in this
session. Prize committee chairs will say a few words about the winning selections.
341
TC69
■ TC69
Maintenance and Renewals. Since railways need to make use as much as possible
of the available data for optimizing their maintenance programs, the required
analytics to support key decisions in an efficient and effective manner will be
illustrated and discussed.
69-Room 201C, CC
Multimodal Traffic Signal Control in a Connected
Vehicle Environment
3 - Using Data Visualization to Assess Performance Risk
Eric Pachman, Director, Network Modeling & Analytics, CSX, 500
Water Street, Jacksonville, FL, 32202, United States of America,
Eric_Pachman@csx.com
Sponsor: TSL/Intelligent Transportation Systems (ITS)
Sponsored Session
At CSX, the way we think about “capacity” is changing. By adding data
visualization to traditional industry modeling tools, discussions on capacity are
shifting to discussions on risk and reliability. Our evolution in capacity analysis is
helping CSX better prioritize infrastructure projects to improve network fluidity.
In addition, through data visualization, we can start to “see” how various
operating and commercial requirements and initiatives impact line of road
capacity and risk.
Chair: K. Larry Head, University of Arizona, Tucson, AZ,
United States of America, larry@sie.arizona.edu
1 - The Multi Modal Intelligent Traffic Signal Control
System (MMITSS)
K. Larry Head, University of Arizona, Tucson, AZ, United States of
America, larry@sie.arizona.edu, Yiheng Feng, Mehdi Zamanipour,
Shayan Khoshmagham, Byunho Beak, Sara Khosravi
4 - Deploying Predictive Analytics Solutions in the Rail Industry and
Seeing a Return on the Investment
Robert Morris, Chief Science Officer, Predikto, Inc., 1320
Ellsworth Industrial Blvd, Suite A1600, Atlanta, GA, 30318,
United States of America, Robert@predikto.com, Mario Montag
The Multi Modal Intelligent Traffic Signal Control System (MMITSS) is a Dynamic
Mobility Application for connected vehicles in signalized networks. MMITSS
provides intelligent signal control, priority control for emergency vehicles, transit,
trucks, and pedestrians, and performance observation. MMITSS has been
implemented in the Arizona Connected Vehicle Testbed in Anthem, AZ.
In this panel, Predikto will provide an overview of automated dynamic predictive
analytics solutions specific to the rail industry. Use cases currently in deployment
across the globe specific to predicting and reducing downtime in freight and
commuter locomotives are discussed alongside the challenges that organizations
face during the process of deploying such technology. Also considered are
strategies to assist in expediting monetary return on investments in predictive
maintenance.
2 - Personalized Signaling for Connected Travelers in a Multi Modal
Traffic Signal System
Sara Khosravi, University of Arizona, Tucson, AZ, United States of
America, sarakhosravi@email.arizona.edu, Sriharsha Mucheli,
K. Larry Head
Smartphones have become standard equipment for almost all travelers. The
smartphone can be used to provide personalized signaling information for multi
modal travelers including pedestrians and bicycles at signalized intersections,
transit riders, and automobile drivers using navigation applications. This talk will
explore how smartphone applications can impact the transportation system.
■ TC71
71-Room 202B, CC
3 - Multi-Modal Intelligent Traffic Signal System, Optimal
Priority Control
Mehdi Zamanipour, University of Arizona, Yiheng Feng,
K. Larry Head, Shayan Khoshmagham
Transportation Planning I
Contributed Session
Chair: Antonio Antunes, Professor, University of Coimbra, Dept. of
Civil Engineering, Coimbra, Portugal, antunes@dec.uc.pt
1 - Plug-in Electric Vehicle Charging Infrastructure Planning using
Cellular Network Data
Jing Dong, Assistant Professor, Iowa State University, 350 Town
Engineering Building, Ames, IA, 50011, United States of America,
jingdong@iastate.edu, Luning Zhang
A priority control algorithm is presented that simultaneously considers the needs
of different modal users in a Connected Vehicle environment. A mathematical
programming framework that allows multiple priority requests to be considered
simultaneously based on a hierarchical control policy at the intersection level will
be presented.
4 - Real-Time Performance Observation under Connected
Vehicle Technology
Shayan Khoshmagham, University of Arizona, K. Larry Head,
Yiheng Feng, Mehdi Zamanipour
This paper presents a method to identify activity-travel patterns, in terms of
timing and duration at home, work, and other major destinations, using multiday
cell phone records. A Hidden Markov Model (HMM) is built to link traveler’s
activity transitions to the observed cell tower locations. The probabilistic
parameters of HMM are estimated using the Baum–Welch algorithm. The derived
travel distances and dwell times are key inputs for plug-in electric vehicle
charging infrastructure planning.
This paper introduces an approach to observe the performance measures of a
multi-modal transportation system in a connected vehicle environment. Different
types of metrics including traffic-based, CV-based and signal-based measures are
observed and estimated by mode by movement. Challenges regarding low market
penetration rate and privacy of the road users are addressed respectively
2 - The Barriers of Electric Vehicles Spread Adoption in China
Faping Wang, Ph.d Candidate, Tsinghua University, Shenzhen
Graduate School of Tsinghua, Shenzhen, GD, 518100, China,
wfp13@mails.tsinghua.edu.cn
■ TC70
70-Room 202A, CC
This paper present a survey research about barriers of electric vehicle spread
adoption in China, 1000 questionnaire was designed and send to participants
which come from Beijing, Shanghai, Guangzhou and Shenzhen, all of which are
large city in China. Majority of participants have EVs driving experience or owner
of EV or PHEV. The demographic data was analyzed by statistic methods which
reveal that more different choice behavior exit between western consumer and
Chinese in EVs consumption.
Predictive Analytics in Railway – Practice
Sponsored Session
Chair: Dharma Acharya, President, KOSU Services LLC, 241
Auburndale Dr., Ponte Vedra, FL, 32081, United States of America,
acharya.dharma@gmail.com
1 - State of Railway Analytics
Dharma Acharya, President, KOSU Services LLC, 241 Auburndale
Dr., Ponte Vedra, FL, 32081, United States of America,
acharya.dharma@gmail.com
3 - Strategic Infrastructure Development for Alternative Fuel Vehicles
with Routing Considerations
Seong Wook Hwang, PhD Student, The Pennsylvania State
University, 232 Leonhard Building, The Pennsylvania State
University, University Park, PA, 16802, United States of America,
soh5223@psu.edu, Sang Jin Kweon, Jose A. Ventura
This research considers decisions on the siting of alternative fuel (AF) refueling
stations and on the routing of AF vehicles when drivers take a detour to refuel.
Supposing that a driver takes any path whose distance is less than or equals to a
detour distance, we provide an algorithm that finds feasible paths between origins
and destinations. Then, a mathematical model is proposed to determine the
locations of AF refueling stations with the objective of maximizing the covered
flows.
A brief overview of how the new emerging technology “Analytics” has been
leveraged by railroads will be presented. Potential new areas where railways
might be able to further utilize this new techniques to bring bottom line value to
the company will also be discussed.
2 - Big Data Analytics for Optimized Track Maintenance and
Renewal Management
Luca Ebreo, MERMEC Inc., 110 Queen Parkway, Columbia, NY,
United States of America, Luca.ebreo@mermecgroup.com,
Pietro Pace
Nowadays, track inspection technology allows railways to collect more and more
data on track’s condition. These data are comparable to “big data” and require
proper analysis in order to extract information for properly managing Track
342
4 - Optimization Model for Floating Carsharing System Planning
Antonio Antunes, Professor, University of Coimbra, Dept. of Civil
Engineering, Coimbra, Portugal, antunes@dec.uc.pt
TC74
1 - Monitoring Uniformity of Particle Distributions in Manufacturing
Processes using the K Function
Xiaohu Huang, Graduate Student, City University of Hong Kong,
106B, Hall 8, Student Residence, CityU, Hong Kong, China,
xhhuang6-c@my.cityu.edu.hk, Qiang Zhou
The focus of this presentation is an optimization model aimed to assist a floating
carsharing company in the making of its key planning decisions – the area to be
operated by the company (called home area), the price rate or rates to be charged
to customers, the relocation operations to perform across zones of the home area,
and, indirectly, the size of the fleet to be used by the company. The results that
can be obtained through the model are exemplified for a real-world setting.
Data in the form of spatial point patterns are frequently encountered in
manufacturing processes. The distributional characteristics of a spatial point
pattern can be summarized by functional profiles like K function. In this study, a
Gaussian process is designed to characterize its behaviour under complete spatial
randomness. A T2 control chart is proposed to monitor the uniformity of point
patterns.
■ TC72
2 - Bayesian Hierarchical Linear Modeling of Profile Data with Apps
to Quallity Control of Nanomanufacturing
Jianguo Wu, Assistant Professor, University of Texas-El Paso, El
Paso, TX, United States of America, jwu2@utep.edu, Yuhang Liu,
Shiyu Zhou
72-Room 203A, CC
Applications III
To achieve a highly automatic quality control, simultaneous profile monitoring
and diagnosis is often required. This paper presents a general framework by using
a hierarchical linear model to connect profiles with both explanatory variables
and intrinsic processing or product parameters for simultaneous monitoring and
diagnosis. The effectiveness is illustrated through numerical studies and
applications to NDE profiles for quality control of nanocomposites manufacturing.
Sponsored Session
3 - Modeling of Optical Profiles in Low-E Glass Manufacturing
Qian Wu, Graduate Student, Texas A&M University, Industrial
and Systems Engineering, College Station, TX, 77843,
United States of America, hi.wuqian@gmail.com, Li Zeng
MS 3131, College Station, TX, United States of America,
1 - The Predict Project: Enhancing DDDAS/Infosymbiotics with
Privacy and Security
Vaidy Sunderam, Emory University, 400 Dowman Dr #W-401,
Math & CS, Atlanta, GA, 30322, United States of America,
vss@emory.edu, Li Xiong
Quality of low-E glass is measured by optical profiles. This study considers
modeling of the optical profile data in Phase I analysis.
4 - Wafer Yield Prediction Based on Virtual
Metrology-generated Parameters
Wan Sik Nam, Seoung Bum Kim/Korea University, Korea
University, 145 Anam-ro, Seongbuk-, Seoul, Korea, Republic of,
wansiknam@korea.ac.kr
The ubiquitousness of mobile devices will greatly expand the applicability of
DDDAS, provided privacy and security issues are addressed. The PREDICT project
is developing: (1) approaches to assign data-targets to participants with privacy
protection; (2) methods for aggregating and fusing data that quantify veracity of
the data sources and maintain high fidelity; and (3) secure distributed
computation for field- and region-level deployment of the DDDAS paradigm with
adaptation and feedback.
Yield prediction is one of the most important issues in semiconductor
manufacturing. Especially, for a fast-changing environment of the semiconductor
industry, accurate and reliable prediction techniques are required. In this study,
we propose a procedure to predict wafer yield using process parameters generated
from the virtual metrology of the semiconductor fabrication, which is based on a
variety of regression and classification algorithms.
2 - Securing Industrial Control Systems with
Software-defined Networking
Dong Jin, Assistant Professor, Illinois Institute of Technology,
10 W 31st Street, Stuart Building 226E, Chicago, IL, 60614,
United States of America, dong.jin@iit.edu
■ TC74
74-Room 204A, CC
Modern industrial control systems (ICSes) are increasingly adopting Internet
technology to boost control efficiency, which unfortunately opens up a new
frontier for cyber-security. With the goal of safely incorporating existing
networking technologies in ICSes, we design a novel software-defined
networking (SDN) architecture for ICSes, with innovative security applications
(e.g., network verification and intrusion detection) and rigorous evaluation using
IIT’s campus microgrid.
Innovative Methods for System Informatics
Sponsored Session
Chair: Peihua Qiu, Professor, University of Florida, 2004 Mowry Road,
Gainesville, FL, 32610, United States of America, pqiu@phhp.ufl.edu
1 - When Importance Sampling Meets Stochastic Simulation Models
Eunshin Byon, Assistant Professor, University of Michigan, 1205
ebyon@umich.edu, Youngjun Choe, Nan Chen
3 - A DDDAS Approach to Distributed Control in Computationally
Constrained Environments (UAV Swarms)
Vijay Gupta, Univ. of Notre Dame, 275 Fitzpatrick Hl Engrng,
Notre Dame, IN, 46556, United States of America,
vgupta2@nd.edu, Greg Madey, Daniel Quevedo, Wann-jiun Ma
In modern applications of distributed control, the traditional assumption of ample
processing power at every time step at each agent can be challenged by use of
processor intensive sensors such as cameras. Inspired by the Dynamic Data Driven
Application System approach, we present an algorithm that shifts computational
loads among the agents to guarantee performance in spite of reduced average
processor availability. Analytical results and numerical simulations illustrate the
approach.
Importance sampling has been used to improve the efficiency of simulations
where the simulation output is uniquely determined, given a fixed input. We
extend the theory of importance sampling to estimate a system’s reliability with
stochastic simulations where a simulator generates stochastic outputs. Given a
budget constraint on total simulation replications, we derive the optimal
importance sampling density that minimize the variance of an estimator.
2 - QQ Models: Joint Modeling for Quantitative and Qualitative
Quality Responses in Manufacturing System
Ran Jin, Virginia Tech., Grado Department of Industrial and,
Systems Engineering, Blacksburg, VA, 24061, United States of
America, jran5@vt.edu, Xinwei Deng
■ TC73
73-Room 203B, CC
A manufacturing system with both quantitative and qualitative (QQ) responses is
widely encountered. The QQ responses are closely associated with each other, but
current methodologies often model them separately. This paper presents a novel
modeling approach, called “QQ models”, to jointly model the QQ responses
through a constrained likelihood estimation. Both simulation studies and a case
study are used to evaluate the performance of the proposed method.
Quality Monitoring and Analysis in Complex
Manufacturing Processes
Sponsored Session
Chair: Li Zeng, Assistant Professor, Texas A&M University, Industrial
and Systems Engineering, College Station, TX, 77843,
United States of America, lizeng@tamu.edu
Co-Chair: Qiang Zhou, Assistant Professor, City Univ of Hong Kong,
Kowloon, Hong Kong, China, q.zhou@cityu.edu.hk
343
TC75
3 - Discussant’s Presentation
Yu Ding, Professor, Texas A&M University, ETB 4016, MS 3131,
College Station, TX, United States of America,
■ TC76
As a discussant in this Technometrics special issue session on system informatics, I
will present my understanding of strengths and weaknesses of the two papers
selected by Technometrics editor for this session. I will also discuss other related
research problems on the similar topics.
Sponsor: Simulation
Sponsored Session
76-Room 204C, CC
Advances in Stochastic Simulation
Chair: Henry Lam, University of Michigan, 1205 Beal Ave., Ann Arbor,
MI, United States of America
1 - Risk Assessment for Input Uncertainty
Helin Zhu, School of Industrial and Systems Engineering, Georgia
Institute of Technology, 755 Ferst Drive NW, Atlanta, GA, 30332,
United States of America, hzhu67@gatech.edu, Enlu Zhou
■ TC75
75-Room 204B, CC
IBM Research Best Student Paper Award III
When simulating a complex stochastic system, the behavior of the output
response depends on the input parameter estimated from finite real-world data,
and the finiteness of data brings input uncertainty to the output response. Risk
assessment for input certainty, which quantifies the extreme behavior of the
mean output response, is extremely important. In the present paper, we introduce
the risk measures for input uncertainty and study the corresponding estimators.
Sponsored Session
huangmh@ntu.edu.tw
huangmh@ntu.edu.tw
2 - Projected Directional Derivatives for High Dimensional
Gradient Estimation
Raghu Pasupathy, Associate Professor, Department of Statistics,
Boqian Zhang
We present a method to estimate gradients in high dimensions by projecting
randomly generated directional derivatives onto the various axes. We discuss
theoretical properties and sampling measures that minimize the resulting
estimator’s error norm. The method appears particularly relevant in high
dimensions since only two observations are needed for a complete gradient
estimator.
2 - Efficient Information Heterogeneity in a Queue
Yang Li,Rotman School of Management, University of Toronto,
105 St. George Street, Toronto ON M5S3E6, Canada,
Yang.Li10@Rotman.Utoronto.Ca, Ming Hu
How would the growing prevalence of real-time delay information affect a service
system? We consider an M/M/1 queueing system in which only a fraction of
customers are informed about real-time delay. Surprisingly, we find that system
throughput and social welfare can be unimodal in the fraction of informed
customers.
3 - Perfect Sampling of GI/GI/C Queues
Yanan Pei, Columbia University, 500 W. 120th St, Mudd 313,
yp2342@columbia.edu, Jose Blanchet, Jing Dong
We introduce the first class of perfect sampling algorithms for the steady-state
distribution of multi-server queues with general inter-arrival time and service
time distributions. Our algorithm is built on the classical dominated coupling from
the past protocol using a coupled multi-server vacation system as the upper
bound process. The algorithm has finite expected termination time with mild
moment assumptions on the inter-arrival time and service time distributions.
3 - Scheduling and Pricing Services for Online Electric Vehicle
Charging
Mark Nejad,Assistant Professor, University of Oklahoma,
Industrial and Systems Engineering, Norman OK, United States of
America, mark.nejad@ou.edu , Ratna Babu Chinnam, Daniel
Grosu, Lena Mashayekhy
4 - Rare Event Simulation in the Neighborhood of a Rest Point
Konstantinos Spiliopoulos, Assistant Professor, Boston University,
Department of Mathematics and Statistics, 111 Cummington
Mall, Boston, MA, 02215, United States of America,
kspiliop@math.bu.edu
We design mechanisms for EV charging services in online settings. We prove that
our proposed mechanisms are incentive compatible, that is, truthful reporting of
price and the amount of charging is a dominant strategy for self-interested EV
drivers. Our preemption-aware charging mechanisms allow providers to manage
fluctuations in renewable energy production.
We construct efficient importance sampling Monte Carlo schemes for finite time
exit probabilities in the presence of rest points. The main novelty of the work is
the inclusion of rest points in the domain of interest. We motivate the
construction of schemes that perform well both asymptotically and nonasymptotically. We concentrate on the regime where the noise is small and the
time horizon is large. Examples and simulation results are provided. Joint work
with Paul Dupuis and Xiang Zhou.
4 - Scheduling with Testing
Thomas Magnanti,Institute Professor, MIT, 77 Massachusetts
Avenue, 32-D784, Cambridge MA 02139, United States of
America, magnanti@mit.edu, Retsef Levi, Yaron Shaposhnik
We study a new class of scheduling problems that captures a common tradeoff
between using resources for processing jobs, and investing resources to ‘test’ jobs
and learn more about their uncertain attributes. This can inform future decisions,
but also delay service. We derive intuitive structural properties of the optimal
policies, and use a new cost-accounting scheme to devise a surprisingly low
dimensional dynamic programming formulation, which ultimately leads to an
FPTAS.
■ TC77
77-Room 300, CC
5 -Trading Time in a Congested Environment
Luyi Yang,Doctoral Student, University of Chicago Booth School
of Business, Chicago, IL, United States of America,
luyi.yang@chicagobooth.edu, Laurens Debo, Varun Gupta
Logistics II
Contributed Session
Chair: Fateme Fotuhiardakani, Data Scientist, TMW Systems, 6085
Parkland Blvd, Mayfield Heights, OH, 44124, United States of America,
fateme.fotuhi@gmail.com
1 - Using Heuristics to Solve the Container Loading Problem
Focusing on Priority Levels and Utilization
Crystal Wilson, Clemson University, 6 Natalie Ct., Greer, SC,
29651, United States of America, crysta3@clemson.edu,
Mary Beth Kurz
We propose a time-trading mechanism, mediated by a revenue maximizing
broker, in which customers privately informed about their waiting costs mutually
agree on the ordering in a queue via trading positions. To that end, we show that
the broker can implement an auction with a trade-participation fee and two trade
restriction prices on customer bids. Under the optimal auction, there is partial
pooling in the bidding strategies and therefore customers are not strictly
prioritized.
Just-in-time manufacturers need the parts to arrive to the facility by a scheduled
time to keep the assembly line moving smoothly. How small containers, such as
parts, are loaded onto a larger container is a special type of packing problem. This
research will focus on creating a heuristic that creates loading patterns that
balances priority levels, while also maximizing the utilization of the container
with respect to the weight and cube.
344
2 - Bundling and Pricing Truckload and Less-than-truckload Services
with Stochastic Demand
Rodrigo Mesa Arango, Assistant Professor, Florida Institute of
Technology, 150 W. University Blvd, Melbourne, FL, 32901,
United States of America, rmesaara@purdue.edu, Satish V.
Ukkusuri
TC79
2 - What Short-term Market Design for Efficient Flexibility
Management in Gas Systems?
Florian Perrotton, IFP Energies Nouvelles / GDF SUEZ /
Université Paris Ouest Nanterre, 7 Rue de Rochechouart, Paris,
75009, France, florian.perrotton@gmail.com
With the increase of electric intermittent renewables, often backed-up by CCGTs,
variability has been transferred to the gas network. Balancing the network has
become a technical and economic issue. Applied to gas systems, market designs
similar to locational marginal pricing might improve the situation. However, in
such markets, flexibility can be handled in different ways. Using a linearized
transient model of the gas network (LP), we analyze the efficiency of two
different auction designs.
Algorithms to bundle and price Truckload (TL) and Less-than-truckload (LTL)
services are proposed considering stochastic demand, value-based pricing, and
demand segmentation. A two-stage min-cost flow problem accounts for uncertain
demand. Its deterministic equivalent is formulated as a regular min-cost flow
problem and efficiently solved. Deterministic models overestimate benefits.
Numerical experiments reveal the cost of uncertainty and demonstrate
improvements in bundle quality.
3 - Optimal Decision Support System for Large Scale Energy
Management in Smart Grid
Sunil Vuppala, IIITB, Electronics City, Bangalore, India,
sunil.vuppala@iiitb.ac.in, Prasanna Gns
3 - Stochastic Models for Optimal Reusable Pallets Management
and Sustainable Supply Chain Configuration
Xiangxiang Fan, Huazhong University of Science and Technology,
Wuhan, China, goodman432700@163.com, Yeming Gong,
Xianhao Xu
We propose a decision support system with modeling methodologies to handle
large scale energy management problems in smart grid. The convexified model is
solved in rolling horizon approach using solvers (Cplex) both at consumer and
utility levels. The systems comes with heuristic for quick run times. We analyze
the stability and accuracy of the system for millions of users with sensitivity
analysis. The results indicate that our system is beneficial to both consumers and
utility companies.
This paper compares the impact of supply chain configuration and operating rules
on the management of reusable pallets in closed-loop supply chains. We evaluate
how pallet management strategy could affect the whole performance of a supply
chain in terms of minimizing the number of pallets and the cycle time and
propose semi-open queuing network models to analysis four different alternatives
which could provide an effective tool for optimal pallet management.
4 - Asset Management Strategies for Wind Turbines
Suna Cinar, Wichita State University, 1845 Fairmount Street,
Wichita, KS, 67260, United States of America,
sxcinar@wichita.edu, Bayram Yildirim
4 - Iterative Mechanisms for Shippers’ Collaboration in ProductionShipping Schedule Planning
Minghui Lai, Assistant Professor, Southeast University, Economics
and Management Building, Jiulonghu Campus, Nanjing City,
China, laimh@seu.edu.cn
The decision to retrofit or keep an aged Wind Turbine (WT) can be a difficult
decision. Based on existing operation and maintenance cost or budget allocated,
one can either retrofit a WT or maintained the existing WT with a preventive
maintenance.The difficult question is which of these options is superior. In this
study, a mixed integer linear programming (MILP) modelling approach is
proposed to determine the trade-off between these two options.
We study the collaborative distribution problem of the shippers with sensitive
private information. We propose iterative mechanisms that are convergent,
strategy-proof, individually rational, and budget balanced in most cases. The
mechanisms are implemented by efficiently computable distributive algorithms.
Extensive simulations show that the mechanisms converges fast and have high
efficiency.
■ TC79
5 - Dynamic Capacitated Intermodal Terminal Location Problem
Fateme Fotuhiardakani, Data Scientist, TMW Systems, 6085
Parkland Blvd, Mayfield Heights, OH, 44124, United States of
America, fateme.fotuhi@gmail.com, Nathan Huynh
79-Room 302, CC
Invited Session
In this paper we introduce a dynamic intermodal terminal location problem
which intends to locate new intermodal terminals to dynamically meet
fluctuations in customers’ demands. The planning horizon is partitioned into
multiple consecutive time periods with a given budget for network configuration
in each time period. A lagrangian relaxation approach embedded with a heuristic
algorithm for upper bound computation is developed to solve this NP-hard
problem.
1 - Syncopation Software - DPL Portfolio and DPMXTM : A Decision
Analysis based System for Better Portfolio Decisions
Chris Dalton, Syncopation Software, 6 State Street, Suite 308,
Bangor, ME, 04401, United States of America,
cdalton@syncopation.com
This demonstration will show how the DPMXô System can serve as the analytical
backbone for an effective portfolio analysis process. We’ll start with an overview
of DPL Professional, a proven modelling environment for decision analysis, risk
analysis and Monte Carlo simulation. Next, we’ll cover DPL Portfolio, the
modelling environment for portfolio analysis, visualization and prioritization.
Finally we’ll show DPMX, a web-based based system for managing project data
and presenting portfolio results in attractive, management-friendly charts. The
motivating examples will be drawn from the prioritization of an R&D portfolio in
the pharmaceutical industry.
■ TC78
78-Room 301, CC
Optimization of Energy Systems
Contributed Session
Chair: Florian Perrotton, IFP Energies Nouvelles & GDF Suez, 228-232
avenue NapolÈon Bonaparte, Rueil-Malmaison, France,
florianperrotton@gmail.com
1 - Distributed Algorithms for the DC-OPF Problem
Hesam Ahmadi, Student, Pennsylvania State University, 107
Holerman Hall, University Park, PA, 16802, United States of
America, ahmadi.hesam@gmail.com, Uday Shanbhag
2 - Frontline Systems, Inc. – Analytic Solver® Platform: Integrated
Data Mining, Simulation and Optimization in Microsoft Excel
Frontline Systems
Analytic Solver Platform in Microsoft Excel has everything you need for
forecasting and data mining, Monte Carlo simulation and risk analysis,
conventional and stochastic optimization – with data from Apache Spark and
visualization of results in Excel, Tableau and Power BI. See how you can use it to
build your own analytic expertise and teach others, leveraging what you already
know, build and solve industrial-scale models with the world’s best Solvers, and
effectively communicate business results.
In this talk, an ADMM-based distributed algorithm is presented for the solution of
the Direct Current Optimal Power Flow (DC-OPF) problem. We consider a dual
formulation that respects privacy concerns and leverage the structure to develop a
distributed scheme. We show that the resulting sequence of primal and dual
iterates is provably convergent. Preliminary numerics are provided.
345
TD01
■ TD02
Military Applications
■ TD01
Contributed Session
Chair: Irene Gerlovin, PhD Candidate/ Part Time Lecturer, Rutgers
Business School, 1 Washington Pl, Newark, NJ, 07102, United States of
America, irene.gerlovin@gmail.com
1 - Modeling Disease Mortality in The National Operational
Environment Model (NOEM)
Venkat Venkateswaran, Rensselaer Polytechnic Institute,
275 Windsor St., Hartford, CT, 06033, United States of America,
venkav3@rpi.edu, John Salerno
Optimizing Decisions in Conflict, Deterrence,
and Peace
Sponsored Session
Chair: Brian Lunday, Assistant Professor Of Operations Research,
Department of Operational Sciences, Grad. Sch. of Engr. & Mgmt.,
Air Force Institute of Technology, Wright Patterson AFB, OH, 45433,
United States of America, Brian.Lunday@afit.edu
1 - Active Target Defense Cooperative Differential Game
David Casbeer, Dr., Air Force Research Laboratory, 2210 8th
Street, B20146 R300, Wright Patterson AFB, OH, 45433,
United States of America, david.casbeer@us.af.mil, Meir Pachter,
Eloy Garcia
The National Operational Environment Model (NOEM) is a large scale stochastic
model that can be used to simulate the operational environment of a nation-state.
Effects of various action alternatives can then be studied through simulations. In
this work we describe the methodology developed to estimate disease mortality.
Extensive V&V tests show that estimated disease death rates compare well with
published values, year by year, for several countries tested.
2 - Optimization of The Canadian Armed Forces Domestic
Transportation Network
Raman Pall, Defence Scientist, Department of National Defence,
1600 Star Top Road, Ottawa, ON, K1B 3W6, Canada,
raman.pall@forces.gc.ca, Abdeslem Boukhtouta
This work addresses an active target defense differential game where an Attacker
pursues a Target. The Target cooperatively teams with a Defender, to maximize
the distance between the Target and the point where the Attacker is intercepted
by the Defender, while the Attacker tries to minimize said distance. The solution
to this differential game provides the min-max optimal heading angles for the
Target and the Defender team, as well as the Attacker.
The Canadian Armed Forces domestic transportation network transports goods
between military bases and depots throughout Canada using a combination of
military transport assets and commercial carriers. In this presentation, we provide
an overview of the network, describing it as a directed graph and analyzing its
efficiency. Recommendations are made on how utilization of the military
resources can be maximized through improvements to the route scheduling.
2 - Approximate Dynamic Programming for the Military Inventory
Routing Problem with Direct Delivery
Matthew Robbins, Assistant Professor Of Operations Research,
Department of Operational Sciences, Grad. Sch. of Engr. &
Mgmt., Air Force Institute of Technology, Wright Patterson AFB,
OH, 45433, United States of America, matthew.robbins@afit.edu,
Brian Lunday, Ian Mccormack, Rebeka Mckenna
3 - Supply Chain Program Management (SCPM) to the Rescue!
F-35 Program
Irene Gerlovin, PhD Candidate/ Part Time Lecturer, Rutgers
Business School, 1 Washington Pl, Newark, NJ, 07102, United
States of America, irene.gerlovin@gmail.com, Yao Zhao
The military inventory routing problem (IRP) with direct delivery is formulated to
model resupply decisions concerning a set of geographically dispersed brigade
combat team elements operating in an austere combat situation. We construct a
Markov decision process model of the military IRP and obtain solutions via
approximate dynamic programming. Designed computer experiments are
conducted to determine how problem features and algorithmic features affect the
solution quality of our policies.
F-35 program had a number of technical challenges. Since its inception in 2001,
the program is seven years behind the schedule and 70% over initial budget. We
review its key SCPM practices to identify root causes for the delays and to
enhance the chance of success for future DOD acquisitions.
3 - Improving Chemotherapy Delivery through the Simulation of
Scheduling Heuristics
Ryan Slocum, Instructor, Department of Mathematical Sciences,
Building 601, United States Military Academy, West Point, NY,
10996, United States of America, ryan.slocum@usma.edu,
Javad Taheri, Thom Hodgson
■ TD03
Inventory Management II
Contributed Session
In the last decade, chemotherapy delivery has largely become an outpatient
service. This has challenged clinics to administer complex treatments to as many
patients as possible within a fixed period of time. We apply selected scheduling
heuristics to reduce patient waiting times and minimize nurse overtime hours. We
present the results of a case study for which our heuristics found two solutions
that respectively reduce the average patient’s waiting time by 20% and annual
overtime by 60%.
Chair: Ruiqi Hou, University of Science and Technology of China, East
Campus USTC, No. 96 Jinzhai Road, Room 367-414, Hefei, 230026,
China, qiqimath@gmail.com
1 - A Continuous Formulation for a Location-Inventory Problem
Considering Demand Uncertainty
Matias Schuster Puga, Université Catholique de Louvain,
Chaussée de binche,151, Mons, 7000, Belgium,
matias.schuster@uclouvain.be, Jean-sébastien Tancrez
4 - A Game Theoretic Model for the Optimal Disposition of Integrated
Air Defense Missile Batteries
Brian Lunday, Assistant Professor Of Operations Research,
Department of Operational Sciences, Grad. Sch. of Engr. &
Mgmt., Air Force Institute of Technology, Wright Patterson AFB,
OH, 45433, United States of America, Brian.Lunday@afit.edu,
Chan Han, Matthew Robbins
We propose a location-inventory model that can be applied to design large supply
chain networks. We address a continuous non-linear formulation that minimizes
transportation, inventory, order, safety stock and facility opening costs. We solve
the non-linear model with an heuristic algorithm that relies on the fact that the
model simplifies to a continuous linear program when two auxiliary variables are
fixed. We show the efficiency of the algorithm with the computation of numerical
experiments.
We examine the allocation of air defense batteries to protect a country’s
population as a three-stage sequential, perfect information, zero-sum game
between two opponents. We formulate a trilevel nonlinear integer program, but
instead apply both an enumeration algorithm and a customized heuristic to
search the game tree. We test both on small instances to assess the efficacy of the
heuristic, and we demonstrate the computational efficiency of the heuristic on
realistic-sized instances.
2 - SQRTN and Portfolio Effect Inventory Models: Notes on Practical
Use and Accuracy for Practitioners
Tan Miller, Director Global Supply Chain Management Program,
Rider University, 12 Winding Way, Morris Plains, NJ, 07950,
United States of America, tanjean@verizon.net, Renato De Matta,
Minghong Xu
We conduct simulations of alternative logistics network inventory stocking
strategies. We then evaluate the accuracy and practical utility to network planners
of using multiple portfolio effect models and the SQRTN model to predict changes
in inventory investment requirements under alternative inventory network
strategies and configurations.
346
3 - A Critique of Empirical Tests on the Inventory-performance
Relationship in U.S. Manufacturing
Daesung Ha, Professor, Marshall University, 415 Corbly Hall,
1 John Marshall Drive, Huntington, WV, 25755,
United States of America, ha@marshall.edu
TD05
4 - An Optimization Model for Allocation and Routing of Municipal
Solid Waste in Flanders
Jens Van Engeland, KU Leuven Campus Brussels, Warmoesberg
26, Brussels, 1000, Belgium, jens.vanengeland@kuleuven.be
In this study, we discuss the estimation errors and the model specification
problem of the existing studies which investigated the relationship between
inventory productivity and firm performance in U.S. manufacturing industry.
Using the panel data of U.S. manufacturing firms over the period of 1980–2014
collected from the Compustat database, we provide the corrected estimation
results.
Historical evolutions and inter-municipal cooperations decide on the current
allocation of municipal solid waste (MSW) to treatment facilities in Flanders.
However, in the near future the region will be confronted with a number of
important waste dilemmas. Therefore it is important to know what gains could be
achieved by redesigning the current allocation in the first place. The proposed
model optimizes the allocation and routing of MSW from municipalities to
treatment plants.
4 - The Effect on Inventories Assets Turnover Change Ratio
by Firm Characteristics
Jihye Lee, Kyungpook National University, Sangyeok
3-Don Buk-Gu, Daegu, Korea, Daegu, Korea, Republic of,
jj2083@gmail.com, Pansoo Kim
5 - A Network Formulation of Competing Demands for Water:
Transaction Costs, Property Rights, and Rents
Patrick O’Reilly, PhD Candidate, Mineral And Energy Economics,
Colorado School of Mines, P.O. Box 11, Golden, CO, 80402,
United States of America, poreilly@mines.edu
This study analyzed the effect on ratio of change of inventories turnover by firm
characteristics using panal data targeting manufacturing companies listed on the
Korea Stock Exchange securities market since Januaru 1, 1999 to December 31,
2012. Firm size, sales growth rate, ROA(return on assets), leverage ratio, credit
rating, age of firm were used as a 6 important firm characteristic variables.
Markets and centrally-planned regimes may be seen as having network structure,
exhibiting not only spatial dependence, but transaction costs and the dual notion
of economic property rights. Water allocation problems pose a range of
institutional questions that network models may be uniquely suited to answer.
This paper investigates transaction cost and economic rent consequences of
choosing between market and command-oriented institutions in light of their
respective network structure.
5 - Service Management in Dynamic Online Markets with Positive
and Negative Word of Mouth
Ruiqi Hou, University of Science and Technology of China,
East Campus USTC, No. 96 Jinzhai Road, Room 367-414, Hefei,
230026, China, qiqimath@gmail.com
■ TD05
We consider that comments online may lead to customers’ leaving the market.
We use the effect that value for money level may have on market size to measure
the economic effect. The model considers both single and two-firm model and the
decision is setting investment cost. Customers are distinguished by their types
which induce heterogeneous rates of adoption information. The information of
value for money level diffuses and affects the transitions of consumers. We
establish conditions for a Nash equilibrium policy.
Social Media Engagement
Invited Session
Chair: Les Servi, The MITRE Corporation, 202 Burlington Road,
Bedford, MA, United States of America, lservi@mitre.org
1 - Development and Evaluation of Tagalog LIWC Dictionaries for
Negative and Positive Emotion
Amanda Andrei, Graduate Student, Georgetown University,
Washington, DC, United States of America
aa1436@georgetown.edu
■ TD04
Economics I
Contributed Session
Developing non-English sentiment analysis tools can ensure that data is not lost
due to language. A proof-of-concept Tagalog Linguistic Inquiry and Word Count
(LIWC) dictionary for positive and negative emotion was developed for use in
analyzing mixed language Twitter data from the Philippines and evaluated against
human-annotated sentiment for Twitter, referred to as groundtruth.
Chair: Patrick O’Reilly, PhD Candidate, Mineral And Energy
Economics, Colorado School of Mines, P.O. Box 11, Golden, CO, 80402,
United States of America, poreilly@mines.edu
1 - Foreign Direct Investment and Organized Crime in Mexico:
A Spatial Approach
Lorena Berumen, Head Of Academic Area In Operations
Management, Universidad Panamericana, Augusto Rodin 498,
Ciudad de México, Mexico, laberumen@up.edu.mx, Roldán
Andrés-rosales, Margarita Hurtado
2 - Consumer Engagement with Green Brands on Facebook as
Revealed in Refined Sentiment Analysis
Tiffany Ting-Yu Wang, Associate Professor, College of
Informatics,KNU, 70-7 Xianyan Rd. Lane 16, Taipei, No, 11688,
Taiwan - ROC, tiffanyt.wang122@gmail.com
Foreign Direct Investment (FDI) has played an important role in the growth and
development of the Mexican economy. In this paper our main contribution is the
analysis of FDI by sector and its spillover effect in the different regions in which
FDI has ben concentrated. Using spatial panel data and a spatial Durbin Model to
assess the direct and indirect effects of FDI on the sectors affected by organized
crime.
The fast-growing number of social media users together with inundating user
generated content has posed significant challenges to firms trying to detect
consumers’ attitudes. This research aims to uncover consumer experiences with
two green brands in the cosmetics industry through collecting and analyzing
public Facebook posts. We refine sentiment analysis by applying the Tetraclasse
model to identify social media context and/or green product attributes as
satisfaction determinants.
2 - The Role of Social Planner in Closed-loop Supply Chain
Lan Wang, Assistant Professor, California State University at East
Bay, 25800 Carlos Bee Blvd., Hayward, CA, 94401, United States
of America, lan.wang@csueastbay.edu, Tharanga Rajapakshe,
Asoo Vakharia
3 - Tailored Incentives and Least Cost Influence Maximization on
Social Networks
Rui Zhang, University of Maryland, College Park, MD, United
States of America, ruizhang@rhsmith.umd.edu, S. Raghavan
We wish to promote a product over a social network and attain 100% adoption.
We study a cost minimization problem where incentives can be tailored to each
individual. A totally unimodular formulation is proposed for trees. Observing that
the influence propagation network is acyclic, we apply this formulation (along
with an exponential set of anti-cycle inequalities) to general networks. Next, a
branch-and-cut approach is developed and used to solve problems on real-world
graphs with 5000 nodes.
Our paper studies the problem of legislation practices on who should be
responsible for recycling, and compares the existing mechanisms on the efficiency
of environmental protection. Given different social objectives – prioritized
consumer welfare, prioritized environmental benefit, or jointly social objective,
we aim to provide roadmap to the social planner on legislation and incentives for
remanufacturing and the end-of-life/use product recycling activities.
3 - Modelling Heterogeneous Economies – Two Competing
Paradigms
Grzegorz Koloch, Warsaw School of Economics, Al.
Niepodleglosci 162, Warsaw, 02-554, Poland,
gkoloch@gmail.com, Mateusz Zbikowski, Bogumil Kaminski
4 - Large-scale Bid Optimization in Online Advertising Auctions
Mustafa Sahin, University of Maryland, Van Munching Hall 3330,
College Park, MD, 20742, United States of America,
mustafa.sahin@rhsmith.umd.edu, Abhishek Pani, S. Raghavan
In sponsored search ads, the search operator collects bids for a given keyword and
determines whose ads will be displayed in what position. The advertisers have to
decide on keywords and positions to bid on given a budget, which can be
modeled as a Multiple Choice Knapsack Problem. The number of keywords and
positions considered can be in the order of hundreds of millions for this
application. We offer an algorithm that is both time and memory efficient and
present results on hard instances.
Two modelling paradigms gained most popularity in the field of heterogeneous
agent macroeconomic modelling: the heterogeneous agent DSGE models and
Agent Based Macroeconomic simulation models. The first approach, is based on
neoclassical foundations and uses dynamic programming paradigm. It is
considered to be a mainstream. The second one still is not used to a comparable
extent both by researchers and policy makers. In this paper we propose an
explanation of the reasons for such a situation.
347
TD06
■ TD06
in the rational expected utility framework.
3 - Sequential Monte Carlo with Parameter Learning for
Long-memory Processes
Konstantinos Spiliopoulos, Assistant Professor, Boston University,
Department of Mathematics and Statistics, 111 Cummington
Mall, Boston, MA, 02215, United States of America,
kspiliop@math.bu.edu
Quantitative Finance and Risk Management
Sponsored Session
Chair: Ning Cai, Associate Professor, Hong Kong University of Science
and Technology, Clear Water Bay, Kowloon, Hong Kong, China,
ningcai@ust.hk
1 - Pricing Asian Options under Markov Processes
Ning Cai, Associate Professor, Hong Kong University of Science
and Technology, Clear Water Bay, Kowloon, Hong Kong, China,
ningcai@ust.hk
We consider state-space models specified up to an unknown vector of parameters
and in which the unobserved state process exhibits long-memory. We estimate
both the state process and the parameter vector and propose a sequential Monte
Carlo method that is based on smoothing of the sample points of model
parameters. We establish a central limit theorem for the state and parameter filter.
We apply the approach to S&P 500 data in the context of a stochastic volatility
model with long memory.
4 - Leveraged ETF Portfolios with Delta-vega Control
Zheng Wang, Columbia University, 116th Street, New York, NY,
10027, United States of America, zw2192@columbia.edu,
Tim Leung
We derive analytical approximations to both continuous and discrete Asian option
prices under general Markov processes. Numerical results illustrate that our
pricing methods are accurate and fast under diffusion models, jump diffusion
models, and pure jump models.
We analyze a collection of static portfolio strategies that allow an investor to
control portfolio sensitivity with respect to the short-term return and realized
volatility of a reference asset. This is done by choosing appropriate weights of
each constituent in a portfolio of leveraged ETFs. We backtest our proposed
strategies using empirical data of major equity leveraged ETFs and illustrate the
efficacy of our methodology.
2 - Transform Methods for Default Timing Problems
Alex Shkolnik, University of California, Berkeley, CA,
United States of America, ads2@berkeley.edu, Kay Giesecke
Reduced-form models of name-by-name default timing are widely used to
measure portfolio credit risk. Combinatorial aspects of many default timing
problems render them NP-complete. Nevertheless, well designed transform
methods do yield efficient algorithms. We illustrate such an algorithm on an
application of CDO pricing. The proposed method reduces computational
complexity by orders of magnitude over those encountered in the literature.
A complete error analysis is provided.
■ TD08
3 - Closed-Form Valuation of Barrier Options
Haohong Lin, Department of Industrial Engineering and Logistics
Management, HKUST, Hong Kong, Hong Kong - PRC,
hlinaa@ust.hk, Ning Cai
Tutorial in Financial Services
Sponsored Session
We study the pricing problem of barrier options that are among the most popular
exotic options in the financial market and derive closed-form pricing formulas
under some option pricing models. Numerical results suggest that our pricing
method is accurate and efficient.
Chair: Bo Zhang, IBM Research, 1101 Kitchawan Road, Route 134,
Yorktown Heights, NY, 10594, United States of America,
zhangbo@us.ibm.com
1 - Reduced Form and Structural Models in Energy Finance
Stathis Tompaidis, Professor, University of Texas at Austin, Office
of Financial Research, Austin, TX, 78712, United States of
America, Stathis.Tompaidis@mccombs.utexas.edu
4 - Does the Prohibition of Trade-throughs Hurt
Liquidity Demanders?
Ningyuan Chen, Columbia University, S. W. Mudd 321, 500 W
120th Street, New York, NY, 10027, United States of America,
nc2462@columbia.edu, Steven Kou
We present both reduced form and structural models used in Energy Finance. The
models span the oil, gasoline, refinery, natural gas, and electricity markets, and
can be used to value generators, oil and natural gas fields, and electricity
generators.
We study the impact of prohibiting trade-throughs on liquidity demanders. We
find that after trade-throughs are prohibited, the transactions of a liquidity
demander might have higher execution cost and effective spread. However, the
additional cost is insignificant for small trades and stocks with abundant liquidity
provision. Our results favor the enforcement of the Order Protection Rule, as the
cost it incurs on liquidity demanders may be outweighed by its benefit.
■ TD09
■ TD07
Collaborative R&D
Sponsored Session
Topics in Optimal Investment
Chair: Niyazi Taneri, SUTD, 8 Somapah Rd, Singapore, Singapore,
niyazitaneri@sutd.edu.sg
1 - Incentivizing External Experts in New Product Development
Shantanu Bhattacharya, Singapore Management University,
Lee Kong Chain School of Business, Grange Heights, Singapore,
238145, Singapore, shantanub@smu.edu.sg, Sameer Hasija
Invited Session
Chair: Mykhaylo Shkolnikov, Princeton University, ORFE, Princeton,
NJ, 08540, United States of America, mshkolni@gmail.com
1 - Arbitrage-free Valuation and Hedging of XVA
Maxim Bichuch, Johns Hopkins University, Baltimore, MD,
United States of America, mbichuch@jhu.edu, Agostino Capponi,
Stephan Sturm
We create a model of new product development where information on external
factors like market potential and technology feasibility is sought from external
experts. The firm has to adequately incentivize these experts to truthfully reveal
their judgment. Contracts are presented to alleviate the resulting adverse selection
problem.
We introduce a framework for computing the Total Valuation Adjustment (XVA)
of an European claim accounting for funding costs, counterparty risk, and
collateral mitigation. We derive the nonlinear BSDEs associated with the
replicating portfolios of long and short positions, and define the buyer and seller’s
XVAs. When borrowing and lending rates coincide we provide a fully explicit
expression for the XVA. When they differ, we derive the semi-linear PDEs, and
conduct a numerical analysis.
2 - Supplier Incentives in Collaborative Product Development with
Internal Competition
Timofey Shalpegin, Lecturer, University of Auckland, 12 Grafton
Road, Auckland, 1010, New Zealand, t.shalpegin@auckland.ac.nz
Internal competition in new product development has a profound, yet
unexplored effect on the incentives of the suppliers involved in a development
project through collaboration with the manufacturer’s competing development
teams. We study the optimal assignment of development teams to different
suppliers. We find that due to the effect of competition on supplier incentives, the
manufacturer may find it optimal to allocate more development teams to a
supplier with lower capabilities.
2 - Rationalizing Behavioral Portfolio Choice
Stephan Sturm, Worcester Polytechnic Institute, Worcester, MA,
United States of America, ssturm@wpi.edu, Carole Bernard
Classical portfolio optimization theory postulates that investors’ preferences are
rational and the optimization criterion is expected utility, for some increasing and
concave utility function. This contrasts with with empirical finding of cognitive
psychology. In this talk we try to answer the question if a given behavioral
portfolio choice in a general incomplete semimartingale market can be replicated
348
TD14
■ TD12
3 - Dynamic Delegated Search
Morvarid Rahmani, Assistant Professor, Georgia Tech,
Atlanta, GA, morvarid.rahmani@scheller.gatech.edu,
Karthik Ramachandran
MAS Tutorial: The State of Operations Research in
the US Military: A 75th Anniversary Perspective
Firms often delegate the search for solution of their innovative problems to third
parties (e.g., search for designs, advertisements, executive leaders, etc.) In this
paper we study how the client’s choice of search process (i.e., defined or openended) depends on the strategic behavior of the provider. Taking the client’s and
provider’s perspective, we identify conditions for which a defined search is
preferred to an open-ended search.
Sponsored Session
Chair: Greg Parlier, Past President, MAS of INFORMS, 255 Avian Lane,
Madison, AL, 35758, United States of America, gparlier@knology.net
1 - The State of Operations Research in the U.S. Military:
A 75th Anniversary Perspective
Greg Parlier, Past President, MAS of INFORMS, 255 Avian Lane,
Madison, AL, 35758, United States of America,
gparlier@knology.net
4 - The Impact of Health Information Technology Bundles on Hospital
Performance: An Econometric Study
Aravind Chandrasekaran, Associate Professer, The Ohio State
University, 2100 Neil Avenue, Columbus, OH, 43210, United
States of America, chandrasekaran.24@osu.edu, Luv Sharma
We examine how two HIT bundles: Clinical (used for patient data collection,
diagnosis and treatment) and Augmented Clinical (used for integrating patient
information and decision making) jointly impact operating cost and process
quality. Results suggest complementarities between these bundles with respect to
process quality but not cost. A posthoc analyses offers additional explanation on
the lack of association with cost.
This extended presentation offers perspectives on the past, present, and future of
Operations Research in the US Department of Defense with emphasis on the
Army. The need for a critical review is argued, and a framework for a
comprehensive assessment is developed. Enduring principles are suggested and
new concepts are presented, including both strategic and transformational
analytics.
■ TD10
■ TD14
Platform-Based Markets in the Digital Era
Engineering Systems Applications
Sponsor: E-Business
Sponsored Session
Sponsored Session
Chair: Jason Chan, Assistant Professor, University of Minnesota, 321
19th Ave S, Minneapolis, MN, United States of America,
jchancf@umn.edu
1 - Dynamic Strategies for Successful Online Crowdfunding
Zhuoxin (Allen) Li, Assistant Professor, Boston College,
140 Commonwealth Avenue, Chestnut Hill, MA, 02467,
United States of America, zhuoxin.li@gmail.com, Jason A. Duan
Chair: Honggang Wang, Assistant Professor, Rutgers University,
96 Frelinghuysen Rd, 201 CoRE, Piscataway, NJ, 08854,
United States of America, honggang.w@rutgers.edu
1 - Optimization of Maintenance Planning for Water Distribution
Network under Stochastic Failures
Xin Chen, Assistant Professor, Southern Illinois University,
P.O. Box 1805, Edwardsville, IL, 62034, United States of America,
xchen@siue.edu, Honggang Wang
This paper empirically investigates the dynamics of investors’ backing behaviors in
the presence of network externalities and a finite time window. Model estimation
shows that investors are more likely to back a project that has already attracted a
critical mass of funding. For the same amount of achieved funding, the backing
propensity declines over time.
Cost-effective and preventive maintenance for water distribution networks
(WDN) is essential for sustainable and reliable use of water resources. We develop
mathematical models and apply optimization procedures for optimal WDN
maintenance planning under stochastic failures. We demonstrate the
mathematical models and optimization approach using a regional WDN in a large
U.S. city. We apply genetics algorithms to solve the optimization problem and find
the optimum maintenance plan for the WDN.
2 - The Effect of Disclosing Purchase Information on Review
Helpfulness: Evidence from Amazon.com
Marios Kokkodis, Assistant Professor, Boston College,
34 E 10th, New York, NY, 10009, United States of America,
mkokkodi@stern.nyu.edu
2 - Optimal Development of Shale Gas Field
Honggang Wang, Assistant Professor, Rutgers University,
96 Frelinghuysen Rd, 201 CoRE, Piscataway, NJ, 08854,
United States of America, honggang.w@rutgers.edu
In this work, we study how the introduction of the Verified Purchase (VP) feature
affected review helpfulness on the Amazon platform. We find that all else equal,
`search’ product VP reviews are on average 3.6% more helpful than nonVP
reviews, and `experience’ product VP reviews are 5.6% more helpful than nonVP reviews.
Optimal development of shale gas involves determining the most-productive
fracturing network via hydraulic stimulation processes in shale reservoirs. Shale
gas development problems can be formulated with mixed-integer optimization
models. We apply a simplex interpolation based optimization method to solve
mixed integer optimization problems associated with shale gas production
projects. The optimization performance is demonstrated with the example case of
developing the Barnett shale field.
3 - The Business Value of Recommendations: A Privacy-preserving
Econometric Analysis
Panagiotis Adamopoulos, Doctoral Candidate, New York
University, 44 W 4th St, New York, NY, United States of America,
padamopo@stern.nyu.edu, Alexander Tuzhilin
3 - Resource Abstraction in Planning and Design of Virtual
Data Centers
Dimitri Papadimitriou, Copernicuslaan 50, 2018, Antwerp, 2018,
Belgium, dimitri.papadimitriou@alcatel-lucent.com
We study the effectiveness of different types of mobile recommendations and
their impact on economic demand, using a privacy-preserving econometric
analysis. Our observational data set is based on a real-world mobile recommender
system, which we further supplement with climate, geospatial, and population
and household data. Based on our findings, an increase by 10% in the number of
times a venue is recommended raises the demand by about 6.7%.
Virtual data centers enable flexible allocation of capacity to customer demands by
aggregating physical resources taken out of a subset of data centers (facilities) to
satisfy customer demands. The goal is to determine the capacity to be provisioned
on opened facilities and the assignment that minimize the cost of opening,
supplying demands and connecting each customer to a subset of facilities. We
compare the resulting cost against the corresponding capacitated facility location
problem.
4 - Effect of Valuation Uncertainty on Buyer Indecision and Bidder
Regret in Online Labor Markets
Kevin Hong, Assistant Professor, Arizona State University,
400 E Lemon St, Tempe, AZ, United States of America,
hong@asu.edu, Paul Pavlou, Alvin Zheng
In online labor markets, 60% of projects fail to reach to a contract, and significant
bidder remorse is observed, indicating a waste of time and effort for both buyers
and freelancers.This paper empirically examines how valuation uncertainty measured as bids price dispersion - affects buyer’s contract indecision and bidders’
regret after the buyer awards a contract. We find bids valuation uncertainty
increases both buyer’s contract indecision and bidders’ regret.
349
TD15
■ TD15
3 - Three Newsvendor Models for Capacity Allocation
Sam Choi, Assistant Professor, Shenandoah University, 1460
University Dr., Winchester, VA, 22601, United States of America,
schoi@su.edu
Capacity Management in Healthcare Operations
We propose three newsvendor models to allocate capacity under uncertainty:
inverse newsvendor, sequential newsvendor, and inverse sequential newsvendor
models. The inverse newsvendor model tries to find optimal demand size under
capacitated environment. The sequential newsvendor model deals with optimal
time durations given that demand sizes. Lastly, the inverse sequential newsvendor
model determines optimal demand sizes given that time durations. We apply
three models to healthcare settings.
Sponsored Session
Chair: Sandeep Rath, PhD Candidate, UCLA Anderson, B501 Gold Hall,
UCLA Anderson, Los Angeles, CA, 90024, United States of America,
Sandeep.Rath.1@anderson.ucla.edu
1 - Workforce Optimization with Patient Volume Variations and
Scheduling Pattern Generation for Hospital
Xuanqi Zhang, Philips Research North America, 345 Scarborough
Rd, Briarcliff Manor, NY, 10510, United States of America,
Xuanqi.Zhang@philips.com, Jingyu Zhang
■ TD17
Routing and Multidimensional
Assignment Applications
A two-stage model is created to optimize hospitals’ workforce which directly
affects hospital cost and patients satisfaction. The model uses simulation-based
stochastic optimization and heuristics to reduce staffing cost, avoid understaffing
and improve scheduling efficiency. The two-stage solution mechanism helps
diminish the gap between staffing optimization research and hospital scheduling
practice. Results were delivered to and tested in hospitals.
Sponsored Session
Chair: Jose Walteros, University at Buffalo, 342 Bell Hall, Buffalo, NY,
United States of America, josewalt@buffalo.edu
1 - Gasoline Replenishment and Routing Problem with Variable
Demands and Time Windows
Yan Cheng Hsu, University at Buffalo (SUNY), 412 Bell Hall,
Buffalo, NY, 14228, United States of America, yhsu8@buffalo.edu,
Rajan Batta, Jose Walteros
2 - Integer Programming Model to Solve Bloodmobiles
Routing Problem
Grisselle Centeno, Associate Professor, Univ. of South Florida,
4202 E. Fowler Ave., Tampa, FL, 33620, United States of America,
gcenteno@usf.edu, Serkan Gunpinar
Blood is a scarce and perishable resource. Approximately 80% of the blood
donations are handled remotely via bloodmobiles. Blood center must determine
the number of bloodmobile units to operate, and designate their daily location(s)
to avoid shortfalls. In this study, a vehicle routing problem is developed using IP.
Optimal routing for each bloodmobile is identified using CPLEX solver & column
generation algorithm. Computational results will be discussed.
An iterative procedure is presented to tackle the gasoline replenishment problem
of gas stations and vehicle routing problem. The problem is formulated as
inventory model with a send-back cost due to the gasoline delivery property that
gas stations should accept ordered quantity completely, to find order quantity and
time window of gas stations, minimizing expected total cost, and as MIP model to
resolve vehicle routing problem, maximizing total profit for transporters.
3 - A Binary Integer Programming Approximation for Vaccine
Vial Distribution
Zahra Azadi, Clemson University, 854 Issaqueena Trl. Apt#902,
Central, SC, 29630, United States of America,
zazadi@clemson.edu, Sandra Eksioglu
2 - Solving Multidimensional Assignment Problems with
Combinatorial Decomposable Cost Functions
Hadi Feyzollahi, State University of New York at Buffalo,
316 Bell Hall, Buffalo, NY, United States of America,
hadifeyz@buffalo.edu, Jose Walteros
One of the challenges faced by health care providers is designing an inventory
replenishment policy for vaccines to ensure a successful immunization of patients
while minimizing purchasing, inventory and wastage costs. Wastage incurs when
doses are disposed from opened vials after their safe use time. We propose an (s,
S) policy which determines the vial size, the reorder quantity, and the order up to
point which optimizes system-wide costs.
We consider several variants of the multidimensional assignment problem (MAP)
where the costs of the optimal assignments are calculated by solving
combinatorial optimization problems. We focus our attention on the cases where
the assignments represent optimal TSP tours, spanning trees, and cliques. We
formulate these MAPs as set partitioning problems and solve them via branch and
price. We develop specific algorithms for solving the corresponding subproblems
for each of the aforementioned cases.
■ TD16
3 - Location-capacity-routing Problem of All-electric
Delivery Vehicles
Nan Ding, University of Buffalo, Buffalo, NY, United States of
America, nanding@buffalo.edu, Rajan Batta
Inverse Optimization
Sponsored Session
All-electric truck adoption becomes one of the main addressees of green logistic
activities, with challenges of limited driving range and long charging time to route
these trucks. This problem aims to handle the challenges of congestion and
waiting at the charging stations. A joint location-capacity-routing (LCR) problem
to determine the optimal location and capacity of charging stations is formulated
and a meta-heuristic approach is proposed to solve this LCR problem.
Chair: Daria Terekhov, Concordia University, 1455 De Maisonneuve
Blvd. W., Montreal, Canada, dterekho@encs.concordia.ca
1 - A Goodness-of-fit Measure for Inverse Optimization
Daria Terekhov, Concordia University, 1455 De Maisonneuve
Blvd. W., Montreal, Canada, dterekho@encs.concordia.ca,
Taewoo Lee, Timothy Chan
4 - Large-scale Neighborhood Search for the Multi-dimensional
Assignment Problem
Alla Kammerdiner, New Mexico State University, P.O. Box 30001,
MSC 4230, Las Cruces, NM, 88003, United States of America,
alla@nmsu.edu, Charles Vaughan
Using an analogy between regression and inverse optimization, we develop a
framework for cost function estimation in linear optimization consisting of a
general inverse optimization model and a corresponding goodness-of-fit metric.
We propose several natural specializations of the framework that evaluate
goodness-of-fit in both the space of decisions and objective value.
The multi-dimensional assignment problem is an NP-hard problem in highdimensional combinatorial optimization. This problem arises in the military
surveillance applications (e.g. sensor data fusion and target tracking) and in
healthcare for ranking exposures to falls. We propose and investigate a new largescale search algorithm for this computationally difficult problem. We evaluate the
performance of new algorithm on various instances and compare it to other stateof-the-art procedures.
2 - Inverse Optimization for the Analysis of Competitive Markets
Michael Pavlin, Wilfrid Laurier University, 75 University Ave,
Waterloo, ON, Canada, mpavlin@wlu.ca, John Birge, Ali Hortacsu
We consider use of inverse optimization as an empirical tool for uncovering
unobservable parameters in competitive markets. In particular, we apply these
techniques to the recovery of transportation and production cost parameters in
natural gas markets.
350
■ TD18
TD20
2 - Non-aggressive Adaptive Traffic Routing
Madhushini Narayana Prasad, Graduate Research Assistant,
Cockrell School of Engineering, University of Texas at Austin,
Austin, TX, 78712, United States of America,
madhushini@utexas.edu, Nedialko Dimitrov
Recent Advances in First Order Methods for LargeScale Optimization
Routing a person through a traffic network presents a dilemma to choose
between fixed route which is an easier to navigate route and adaptive route
which minimizes the travel time by adjusting to the traffic conditions. We
investigate methods for non-aggressive, adaptive routing that is middle-ground
seeking the best of both these extremes, i.e. adaptive routes restricted in number
of route shifts allowed at a critical juncture, and investigate the trade-offs
between the extremes.
Invited Session
Chair: Mingyi Hong, Iowa State University, 3015 Black Engineering,
Ames, IA, 50011, United States of America, mingyi@iastate.edu
1 - On the Information-adaptive Variants of the Admm:
An Iteration Complexity Perspective
Shuzhong Zhang, Professor, University of Minnesota, Department
of Industrial and Systems Eng, Minneapolis, MN, 55455, United
States of America, zhangs@umn.edu, Xiang Gao, Bo Jiang
3 - Social Network Echo Chambers and Popularity
Yinhan Liu, University of Texas Austin, 1901 Crossing Place
#3301, Austin, TX, 78741, United States of America,
yinhan.liu@utexas.edu, Nedialko Dimitrov
We present a suite of variants of the ADMM, where the trade-offs between the
required information on the objective and the computational complexity are
explicitly given. The new variants allow the method to be applicable on a much
broader class of problems where only noisy estimations of the gradient or the
function values are accessible, yet the flexibility is achieved without sacrificing the
computational complexity bounds.
Social network users often have the goal of building a large follower base. Some
users are members of what we term echo chambers, a small group of users that
re-share each other’s messages. We present an empirical study on the impact of
echo chambers on the popularity of users using historical data from Twitter.
Specific questions we address are: Does echo chamber membership increase reshares outside the echo chamber? Does echo chamber membership increase
follower base?
2 - An Optimal Randomized Incremental Gradient Method
Guanghui Lan, University of Florida, Gainesville, FL,
United States of America, glan@ise.ufl.edu
■ TD20
We present a randomized incremental gradient method and show that this
algorithm possesses unimprovable rate of convergence for convex optimization.
We provide a natural game theoretic interpretation for this method as well as for
the related Nesterov’s optimal method. We also point out the situations when this
randomized algorithm can significantly outperform the deterministic optimal
method.
Banking and Insurance
Contributed Session
Chair: Linna Du, Data Scientist, CACS, 2259 Adam Clyton Powell,
New York, NY, 10027, United States of America, linna.du@gmail.com
1 - Success Drivers of Online Equity Crowdfunding Campaigns for
Unaccredited Investors
Anna Lukkarinen, Aalto University, P.O. Box 21220, Helsinki,
00076, Finland, anna.lukkarinen@aalto.fi, Jeffrey Teich,
Hannele Wallenius, Jyrki Wallenius
3 - On the Expected Convergence of Randomly Permuted ADMM
Ruoyu Sun, Stanford University, Menlo Park, CA, 94025, United
States of America, sundirac@gmail.com, Zhi-Quan Luo, Yinyu Ye
Recently, it has been shown that the direct extension of the alternating direction
method of multipliers (ADMM) to the multi-block case fails to converge when
solving a simple square system of linear equations. In this paper, however, we
prove that, if in each step one randomly and independently permutes the
updating order of any given number of blocks, the method will converge in
expectation for solving the square system of linear equations.
Using data from a leading equity crowdfunding platform in Northern Europe, we
explore success factors of campaigns. The results suggest that the investment
criteria traditionally used by professional investors are not of prime importance
for success in equity crowdfunding. Instead, success is related to pre-selected
crowdfunding campaign characteristics and networks. Understanding the success
factors of online equity crowdfunding campaigns is important to the design of
online platforms.
4 - Alternating Direction Method of Multipliers for Distributed Sparse
Principal Component Analysis
Davood Hajinezhad, Iowa State University, 62 B Schilletter
village, Ames, IA, 50010, United States of America,
dhaji@iastate.edu, Mingyi Hong
2 - The Mover-Stayer Process for the Credit Data
Anna Matuszyk, Assistant Professor, Warsaw School of
Economics, Niepodleglosci 162, Warsaw, 02-554, Poland,
amatuszyk@matuszyk.com, Halina Frydman
We propose distributed algorithms to perform sparse PCA. They are quite flexible,
in the sense that they are able to handle different forms of data partition (i.e.,
partition across rows or columns of the data matrix). Numerical experiments
based on both real and synthetic data sets, conducted on high performance
computing clusters, demonstrate the effectiveness of our approaches.
Using the credit data set, coming from the European bank, we estimate the
mover-stayer model, which is an extension of the Markov chain. This model
assumes that the population is heterogeneous: there are “stayers” and “movers”.
“Movers” evolve according to a Markov Chain with the one-step transition
matrix, while “stayers” never leave their initial states. The probability of a
customer being a stayer in a paid up state is modeled using the logistic regression.
■ TD19
3 - Monopolistic Dealer Versus Broker: Impact of Proprietary
Trading with Transaction Fees
Yuan Tian, Ryukoku University, 67 Tsukamoto-cho, Fukakusa,
Fushimi-ku, Kyoto, Japan, tian@econ.ryukoku.ac.jp,
Katsumasa Nishide
Network Inference
Sponsored Session
Chair: Nedialko Dimitrov, Assistant Professor, UT Austin, University of
Texas at Austin, Austin, United States of America,
ned@austin.utexas.edu
1 - Fast, Approximate Inference on Graphical Models by
Reducing Treewidth
Areesh Mittal, University of Texas at Austin, 1626 West 6th St.
Apt. F, Austin, TX, 78703, United States of America,
areesh0612@gmail.com, Nedialko Dimitrov
We consider a one-period financial market with a monopolistic dealer/broker and
an infinite number of investors. While the dealer (with proprietary trading)
simultaneously sets both the transaction fee and the asset price, the broker (with
no proprietary trading) sets only the transaction fee, given that the price is
determined according to the market-clearing condition among investors. We
effectively demonstrate how proprietary trading affects market equilibrium and
welfare of investors.
4 - A Data Analytics Based Approach for Building 360 View of
Banking Customer
Tianzhi Zhao, IBM, Diamond Bld, ZGC Software Park, Beijing,
China, zhaotzhi@cn.ibm.com, Zhen Huang, Ming Xie, Bing Shao,
Yuhang Liu, Jian Xu, Wenjun Yin, Yuhui Fu
Complexity of exact inference algorithms in graphical models is exponential in
treewidth. We develop technique to perform approximate inference by removing
edges and updating factors, leading to reduced treewidth. We prove bounds on
error in approximation. Finding updated factors involves solving a geometric
program (GP) with exponential number of constraints. We develop row
generation technique to solve the GP. We demonstrate the results on discrete
graphical models applied to social networks.
Banks today are experiencing transformation from product centricity to customer
centricity. With advent of big data, it enables banks to fully and deeply
understand customers by building 360 degree customer view. In this paper, a data
analytics based approach for 360 degree view of banking customer is proposed. It
can help banks quick build customer centricity for customer segments, targeted
marketing, personalized recommend, etc.
351
TD21
■ TD22
5 - Credit Scoring using Dynamic State Space Model under
Statistical Volatility
Linna Du, Data Scientist, CACS, 2259 Adam Clayton Powell,
linna.du@gmail.com
Contact Centers
Sponsored Session
In emerging market where the credit score and credit history are not trustworthy,
the estimation and prediction of the credit score and prepayment risks are very
important. In the paper, we propose a dynamic state space model considering the
volatility and dynamic feature of the lending market. We found that the time
varying volatility model provides better prediction than other time series models.
We also identify the key factors that drive the lending risks.
Chair: Rouba Ibrahim, University College London, London, N7 8EP,
United Kingdom, rouba.ibrahim@ucl.ac.uk
1 - Telephone Call Centers: Asymptotic Optimality of Myopic
Forecasting-scheduling Scheme
Han Ye, University of Illinois at Urbana Champaign, 350 Wohlers
Hall, 1206 South Sixth Street, Champaign, IL, 61820, United
States of America, hanye@illinois.edu, Noah Gans, Haipeng Shen,
Yong-Pin Zhou
■ TD21
Disease Modeling in OR
We determine workforce schedules for call center arrivals that are doubly
stochastic. Period-by-period arrival rates follow a hidden AR(1) process, and only
arrival counts are observed. We formulate stochastic programs to minimize longrun average staffing costs, subject to a long-run average constraint on
abandonment. We show that, in steady state, repeated, myopic solution of the
single-period problem is stable, has low cost, and meets the abandonment
constraint.
Sponsored Session
Chair: Emine Yaylali, Senior Service Fellow, Centers for Disease Control
and Prevention, 1600 Clifton Road, Atlanta, GA, 30333,
United States of America, wqq3@cdc.gov
1 - The Potential Impact of Reducing Indoor Tanning on Melanoma
Prevention in the United States
Yuanhui Zhang, CDC, Chamblee GA 30341,
United States of America,, yfp5@cdc.gov, Donatus Ekwueme,
Sun Hee Rim, Meg Watson, Gery Guy
2 - A Structural Model for Agents’ Strategic Behavior in Call Centers
Dongyuan Zhan, University of Southern California,
Los Angeles, CA, United States of America,
Dongyuan.Zhan.2015@marshall.usc.edu, Amy Ward,
Seyed Emadi
We do an empirical study of agent behavior in call centers. We begin by observing
that regression analyses have low explanatory power, even though the data
shows that agents speed up or slow down depending on the system load and their
fatigue level. This leads us to investigate utility based structural models for agent
behavior.
More than 700,000 adults in the United States are treated for melanoma each
year, resulting in annual direct medical costs of $3.3 billion dollars and 9,000
deaths. We developed a Markov model to estimate the health and economic
impacts of reducing indoor tanning for melanoma prevention in the United States
under certain assumptions. According to this model, reducing indoor tanning may
result in favorable savings in medical costs and life-years, comparable to other
national prevention efforts.
3 - Capacity Sizing with a Random Number of Agents
Rouba Ibrahim, University College London, London, N7 8EP,
United Kingdom, rouba.ibrahim@ucl.ac.uk
2 - Estimating the Impact of HIV Care Continuum Interventions on the
Reproduction Number
Yao-Hsuan Chen, CDC, Chamblee GA 30341,
United States of America,xhj1@cdc.gov, Andrew Hill,
Paul G. Farnham, Stephanie L. Sansom
We study the problem of staffing many-server queues with general abandonment
and a random number of servers. For example, uncertainty in the number of
servers may arise in virtual call centers where agents are free to set their own
schedules. We rely on a fluid model to determine optimal staffing levels, and
demonstrate the asymptotic accuracy of the fluid prescription. We also
characterize the optimal staffing policy with self-scheduling agents.
We used a compartmental model to study HIV transmission in the United States
from 2006 through 2020 among heterosexuals, men who have sex with men,
including bisexual men, and injection drug users. We analyzed the impact of
interventions to improve HIV diagnosis, care, and treatment on the reproduction
number. Analyses using this model can provide insights into the long-term
effectiveness of HIV prevention strategies.
■ TD23
3 - Stratifying Risk Groups in Compartmental Epidemic Models:
Where to Draw the Line?
Margaret L. Brandeau, Professor, Stanford University, MS&E
Department, Stanford, CA, 94305, United States of America,
brandeau@stanford.edu, Jeremy D. Goldhaber-fiebert
Markov Decision Models and Approximations for
Manufacturing
Invited Session
Disease models used to support cost-effectiveness analyses of health interventions
are often stratified to reflect population heterogeneity (e.g., age, gender, risk
behaviors). We examine the impact of population stratification in dynamic disease
transmission models: specifically, the impact of different divisions of a population
into a low-risk and a high-risk group. We show that the way in which the
population is stratified can significantly affect cost-effectiveness estimates.
Chair: Tugce Martagan, Eindhoven University of Technology, 5600 MB
Eindhoven, Eindhoven, Netherlands, T.G.Martagan@tue.nl
1 - Robust Approximate Dynamic Programming and Structured
Policies for Degradable Energy Storage
Marek Petrik, IBM, 1101 Kitchawan Rd., Yorktown Heights, NY,
10598, United States of America, mpetrik@us.ibm.com
4 - Developing a Dynamic Compartmental Model of HIV in the
United States
Emine Yaylali, Senior Service Fellow, Centers for Disease Control
and Prevention, 1600 Clifton Road, Atlanta, GA, 30333,
United States of America, wqq3@cdc.gov, Paul G. Farnham,
Stephanie L. Sansom, Katherine A. Hicks, Emily L. Tucker,
Amanda Honeycutt
Batteries hold great promise for energy storage in arbitrage in electric grids but
can degrade rapidly with use. In this talk, we analyze the impact of storage
degradation on the structure of optimal policies and describe robust approximate
dynamic programming methods that take advantage of the policy structure.
2 - Component Reservation for Asymptotically Optimal Allocation in
Assemble to Order Production Systems
Haohua Wan, University of Illinois at Urbana-Champaign,
104 South Mathews Ave., Urbana, IL, United States of America,
hwan3@illinois.edu, Qiong Wang
Over 1 million people in the US are living with HIV. To observe trends in HIV and
evaluate the effectiveness of prevention interventions, we developed a dynamic
compartmental model of disease progression and transmission. The population
was stratified by age, sex, circumcision status, race/ethnicity, transmission group,
and risk level. People progressed between compartments defined by disease status
and care and treatment stage. Outcomes included HIV incidence, prevalence, and
care status.
Component reservation is not myopically optimal as it sometimes holds back
components from existing demands. We prove that in many cases, without
reservation, component allocation cannot be asymptotically optimal, i.e., the
percentage difference of the discounted inventory cost from its lower bound does
not converge to zero as demand and production volumes increase, even though
such convergence is achievable under other policies that reserve components for
high-value product demands.
352
TD25
■ TD25
3 - Replenishment and Fulfillment Based Aggregation for General
Assemble-to-Order Systems
Emre Nadar, Assistant Professor, Bilkent University, Department
of Industrial Engineering, Bilkent University, Ankara, Turkey,
emre.nadar@bilkent.edu.tr, Alan Scheller-wolf, Alp Akcay,
Mustafa Akan
Economic Models and Analysis of Networks
and Platforms
Sponsored Session
We present an approximate dynamic programming method to optimizing
Markovian assemble-to-order systems. We alleviate the computational burden by
reducing the large state space of the problem via a novel aggregation method that
builds upon certain component and product characteristics. We show the
optimality of a lattice-dependent base-stock and rationing policy for the aggregate
problem. We also derive finite error bound for the cost function of the aggregate
problem under a mild condition.
Chair: Soumya Sen, University of Minnesota, Minneapolis, MN,
United States of America, ssen@umn.edu
1 - Payments for Transactions Versus Payments for Discoveries:
Theoretical Analyses
Karthik Kannan, Purdue University, 403 W.State Street,
kkarthik@purdue.edu, Rajib Saha
4 - Optimal Manufacturing Policies for Engineer-to-Order Proteins
Tugce Martagan, Eindhoven University of Technology, 5600 MB
Eindhoven, Eindhoven, Netherlands, T.G.Martagan@tue.nl,
Ananth Krishnamurthy
eBay.com in the U.S. charges payments for transactions but Taobao.com in China
charges for discoveries. We theoretically study such payment schemes used by the
platforms. We surprisingly find that when payments are for discoveries, the
platform has an incentive to make welfare-decreasing matches between sellers
and buyers. Similarly, in order for payments for transactions to be sustained,
buyers and sellers should sufficiently value factors such as trust provided by the
platform.
We develop continuous state Markov decision models to optimize design
decisions related to protein purification operations. We focus on engineer-toorder proteins with strict production requirements on quality and yield. We
present a state aggregation mechanism to solve industry-size problems. Our
models and insights are implemented in practice.
2 - The Impact of Online Word of Month on Channel
Disintermediation for Information Goods
Brian Lee, University of Connecticut, 2100 Hillside Road Unit
1041, Storrs, CT, 06268, United States of America,
brian.lee@business.uconn.edu, Xinxin Li
■ TD24
Social Network Analytics
With the advance in digital technology, creators of intellectual products can sell
their work directly to consumers without the help of publishers. In this study, we
construct an analytical model to examine the role of online word of mouth
(eWOM) in this trend of disintermediation. We find that eWOM may encourage
creators to reintermediate publishers for high quality work. Our model makes
predictions on when eWOM benefits publishers and for what types of
products/creators it has the most impact.
Sponsored Session
Chair: Xi Wang, The University of Iowa, S343 PBB, Iowa City, IA,
52242, United States of America, xi-wang-1@uiowa.edu
1 - Optimizing Hurricane Warning Dissemination Problem for
Evacuation Decision Making
Dian Sun, Harbin Engineering University, 172 Princeton Ave.
Apt 1, Buffalo, NY, 14226, United States of America,
sundian@hrbeu.edu.cn, Yan Song, Zifeng Su
3 - Electric Vehicle Power Plants: Carsharing Optimization with
Smart Electricity Markets
Micha Kahlen, Erasmus University Rotterdam, Burgemeester
Oudlaan 50, Rotterdam, Netherlands, kahlen@rsm.nl, Wolf Ketter
Individual make evacuation decisions based on risk perception which can be
socially influenced as evacuation warnings spread through social networks. In this
study a formal model for evacuation warning dissemination in social networks
through time is presented to characterize the social influence of the risk
perception in the evacuation decision making process. Simulation models are
developed to investigate the effects of community mixing patterns and the
strength of ties on evacuation decision.
We study electric vehicles as power plants to bridge weather dependent energy
shortages from wind and solar energy. Particularly, we are interested in the
allocation of electric vehicles by making a trade-off between driving and storing
electricity. This allocation is optimized in a first-price sealed bid auction with
pricing signals from smart electricity markets and the availability of electric
vehicles. Results show positive effects for drivers, carsharing operators, and the
environment.
2 - Inferring User Location from Geographic and Social Network
Da Xu, PhD Student, University of Utah, 1032 E 400 S, Apt 504B,
Salt Lake City, UT, 84102, United States of America,
Da.Xu@business.utah.edu, Xiao Fang
4 - Should You Go with ``Pay as You Go’’?:
Optimal Design of Bucket Plans for Multi-unit Goods
Manish Gangwar, Assistant Professor, ISB, ISB Campus,
Gachibowli, Hyderabad, India, manish_gangwar@isb.edu,
Hemant Bhargava
The lack of tools to monitor the time-resolved locations of individuals constrains
us to gain a deep understanding of human mobility. While, despite the diversity
of people’s travel history, human mobility follows a high of temporal and spatial
regularity. In this paper, we study the human mobility through social and
geographic networks, give a deep insight to how individual mobility pattern and
social network impact with each other, and build a probabilistic model to depict
human mobility.
Among the class of nonlinear tariffs, “Three Part Tariff” is the most general tariff
but it tends to focus on heavy users. Given the evident optimality of a bucket
plan, we ask what are the pros and cons of alternative pricing models? We derive
the closed-form expressions for commonly used demand function and specify a
system of equations with economic interpretation for the general problem. We
also examine the properties of optimal “Three Part Tariff” in the presence of a perunit plan.
3 - Concurrent Diffusions of Information and Behaviors in Online
Social Networks
Shiyao Wang, University of Iowa, 634 Westgate St. Apt. 55,
Iowa City, IA, 52246, United States of America,
shiyao-wang@uiowa.edu, Kang Zhao
Using the spread of the Ice Bucket Challenge (IBC) on Twitter as a case study, this
research compared the concurrent diffusion patterns of both information and
behaviors in online social networks. Individual behaviors related to IBC were
detected by text mining techniques. Comparison between diffusion dynamics of
information and behaviors at different levels revealed interesting differences and
interactions between the two diffusion processes and laid foundations for future
analytics.
353
TD26
■ TD26
2 - Evaluating Lignocellulosic Biomass Supply Chains Considering a
Multi-objective of Optimizing Cost
Burton English, Professor, The University of Tennessee, 2621
Morgan Hall, Knoxville, TN, 37922, United States of America,
benglish@utk.edu, James Larson, Edward Yu, Jia Zhong
Production and Scheduling I
Contributed Session
A switchgrass supply chain that considers the optimization of cost, GHG emissions
and soil erosion for a cellulosic biofuel plant is developed. Using an augmented
epsilon constraint multi-objective optimization model and a compromise solution
method, along with high-resolution spatial data the optimal placement of
feedstock supply chains can be estimated. Spatial characteristics, including land
coverage and infrastructure availability, are crucial to both the cost and the
environmental results.
Chair: Srimathy Mohan, Associate Professor, Arizona State University,
Department of Supply Chain Management, Tempe, AZ,
United States of America, srimathy@asu.edu
1 - Weekly Production Planning on the Basis of Average
Value-at-Risk by Shapley Value
Nobuyuki Ueno, Hiroshima University of Economics, 5-37-1 Gion
Asaminami-ku, Hiroshima, Japan, ueno@pu-hiroshima.ac.jp,
Hiroshi Morita, Koji Okuhara
■ TD28
Under demand uncertainty,they used stock-out ratio for estimating the risk. In
this presentation, we propose a formulation for weekly production planning
problem that reflects the AVaR (Average value-at-risk) for weighing tail risk and a
solution by Shapley value. The characteristics of the solution procedure is proved.
It has the features that it does not require strict probability distribution of stockout and it enables an extension to the case where demand for each period is
correlated.
Dynamic Matching Markets
Cluster: Auctions
Invited Session
Chair: John Dickerson, CMU, 9219 Gates-Hillman Center, Pittsburgh,
PA, 15213, United States of America, dickerson@cs.cmu.edu
1 - Global Kidney Exchange
Afshin Nikzad, Stanford University, 37 Angell Court,
Apt 116, Stanford, CA, 94305, United States of America,
afshin.nikzad@gmail.com, Mohammad Akbarpour, Alvin Roth
2 - A Generalized Dantzig-Wolfe Decomposition Algorithm for Mixed
Integer Programming Problems
Xue Lu, London School of Economics and Political Science,
Houghton Street, London, WC2A 2AE, United Kingdom,
X.Lu7@lse.ac.uk, Zeger Degraeve
We propose a generalized Dantzig-Wolfe decomposition algorithm for mixed
integer programming. By generating copy variables, we can reformulate the
original problem to have a diagonal structure which is amendable to the DantzigWolfe decomposition. We apply the proposed algorithm to multi-level capacitated
lot sizing problem and production routing problem. Rigorous computational
results show that our algorithm provides a tighter bound of the optimal solution
than all the existing methods.
In some countries, many patients die after a few weeks of diagnosis mainly
because the costs of kidney transplantation and dialysis are beyond the reach of
most citizens. We analyze the two proposals in which patients with financial
restrictions who have willing donors participate in kidney exchange without
paying for surgery. Our proposals can save thousands of patients, while
substantially decreasing the average dialysis costs; in particular, we prove that
they are “self-financing”
3 - The Impact of Postponement Practices on the Lot-sizing
Decisions of a Wine Bottling Plant
Sergio Maturana, Professor, Pontificia Universidad Catolica de
Chile, Vicuna Mackenna 4860, Santiago, Chile,
smaturan@ing.puc.cl, Mauricio Varas
2 - Matching with Stochastic Arrival
Neil Thakral, Harvard, 1805 Cambridge Street, Cambridge, MA,
United States of America, nthakral@fas.harvard.edu
We study matching in a dynamic setting, with applications to public-housing
allocation. Objects of different types that arrive stochastically over time must be
allocated to agents in a queue. When objects share priorities over agents, we
propose an efficient, envy-free, and strategy-proof mechanism. The mechanism
continues to satisfy these properties if and only if the priority relations are acyclic.
Estimated welfare gains over existing housing-allocation procedures exceed
$5000 per applicant.
Export-focused wineries face a difficult problem when planning their bottling
lines due to the number of different products they have to bottle and label. A way
of reducing misallocation due to demand variability is by postponing the labeling
process. We propose two MIP planning models that support tactical lot-sizing
decisions. We tested both models in a rolling horizon framework, under different
conditions of capacity tightness, horizon length, and demand uncertainty and we
report the results.
3 - Dynamic Kidney Exchange with Heterogeneous Types
Maximilien Burq, Student, MIT, 77 Massachusetts Avenue,
mburq@mit.edu, Itai Ashlagi, Vahideh Manshadi, Patrick Jaillet
4 - Scheduling of Maximizing Total Job Value with Machine
Availability Constraint
Eun-Seok Kim, Middlesex University, The Burroughs, London,
NW4 4BT, United Kingdom, e.kim@mdx.ac.uk, Joonyup Eun
Kidney exchange programs face growing number of highly sensitized patients. We
develop an online model that models such heterogeneity, and we prove that
having some easy-to-match patients in the pool greatly reduces waiting times
both in the presence of bilateral matching and chain matching. We provide
simulations showing that some prioritizing leads to improved overall efficiency.
We study a single machine scheduling problem of maximizing total job value with
machine availability constraint. The value of each job is given as a non-increasing
step function of its completion time. We develop a branch-and-bound algorithm
and a heuristic algorithm for the problem. Finally, we perform computational
experiments showing that the developed algorithms provide efficient and effective
solutions.
4 - Competing Dynamic Matching Markets
Sanmay Das, WUSTL, One Brookings Dr, CB 1045,
St. Louis, MO, 63130, United States of America,
sanmay@wustl.edu, John Dickerson, Zhuoshu Li,
Tuomas Sandholm
■ TD27
We extend a framework of dynamic matching due to Akbarpour et al. to
characterize outcomes in cases where two rival matching markets compete. One
market matches quickly while the other builds thickness by matching slowly. We
present analytical and simulation results, both in general and for kidney
exchange, demonstrating that rival markets increase overall loss compared to a
single market that builds thickness.
Applications of Multi-objective Optimization
Sponsored Session
Chair: Matthias Ehrgott, Professor, Department: Management Science,
Lancaster University, The Management School, Lancaster, 00, LA1 4YX,
United Kingdom, m.ehrgott@lancaster.ac.uk
1 - A Hybrid Decision Making Approach for Multi-Objective
Infrastructure Planning
Hana Chmielewski, North Carolina State University,
Campus Box 7908, Raleigh, NC, 27695, United States of America,
htchmiel@ncsu.edu, Ranji Ranjithan
A hybrid approach using evolutionary computation and dynamic programming is
used to optimize investments and operational decisions in a water supply case
study system. Solutions are categorized by network centralization metrics, and
analyzed with respect to multiple planning objectives.
354
■ TD29
TD31
3 - A Decision Support System for Traffic Diversion around
Construction Closures
Arezoo Memarian, Graduate Research Assistant, University of
Texas at Arlington, 425 Nedderman Hall, 416 Yates St.,
Box 19308, Arlington, TX, 76019, United States of America,
arezoo.memarian@mavs.uta.edu, Siamak Ardekani
Joint Session Analytics/HAS:The Emerging Role of
Health Systems Engineering and its Impact on
Clinical Informatics and Analytics
The objective of this study is to develop a PC-based decision support tool with a
user-friendly graphical interface to allow development of optimal traffic
management plans around highway construction sites. In addition to the
capability to identify optimum traffic diversion routes, such a tool would also
allow simulation of various traffic management plan scenarios envisioned by
experts.
Sponsor: Analytics
Sponsored Session
Chair: John Zaleski, Chief Informatics Officer, Nuvon, Inc.,
4801 S. Broad Street, Suite 120, Philadelphia, PA, 19112,
United States of America, jzaleski@nuvon.com
1 - How to Make Clinically Actionable Alarms
Jeanne Venella Dnp, Chief Nursing Officer, Nuvon,
4801 S Broad St, Philadelphia, PA, 19112,
United States of America, jvenella@nuvon.com
4 - Can Virtualization Maturity Impact Software Development Project
Performance: An Empirical Study
Rohit Nishant, Assistant Professor, ESC Rennes School of
Business, 2 Rue Robert d’Arbrissel CS 76522, Rennes, 35065,
France, rohit.nishant@esc-rennes.com, Bouchaib Bahli
In this article we invoke IT asset classes’ taxonomy and IT capability maturity
model to examine the impact of virtualization maturity on software development
project performance. Our findings suggest that virtualization capability has a
distinct impact on software development project and process performance.
Finally, this study extends virtualization maturity model’s validity. Implications
for research and practice are discussed.
How to Make Clinically Actionable Alarms The very alarm systems that were
created to enhance patient safety have themselves become an urgent patient
safety concern. We need to fix our current state of alarm systems. We must
achieve both a higher level of sensitivity and specificity. Therefore; reducing both
false and non-actionable alarms. Our goal is ignite the talk on alarm fatigue, begin
to define algorithms for smarter actionable alarms and provide a safer health care
environment.
2 - The Kalman Filter and its Application to Real-time Physiologic
Monitoring of High-acuity Patients
John Zaleski, Chief Informatics Officer, Nuvon, Inc.,
4801 S. Broad Street, Suite 120, Philadelphia, PA, 19112,
United States of America, jzaleski@nuvon.com
■ TD31
Time Series Data Mining
The Kalman Filter (KF) has seen application in many fields owing to its rapid
computational framework and intrinsic optimality in tracking time-series. In this
presentation, the KF is used to optimally track and smooth signal artifact
associated with patient physiologic monitoring.
Sponsored Session
Chair: Mustafa Gokce Baydogan, Assistant Professor, Bogaziai
University, Department of Industrial Engineering, Bebek, Istanbul,
34342, Turkey, mustafa.baydogan@boun.edu.tr
1 - On the Parameter Identification of a New Knot Selection
Procedure in Mars
Cem Iyigun, Associate Professor, Middle East Technical
University, ODTU Kampusu Endustri Muhendisligi Bolum,
Oda 331 Cankaya, Ankara, 06801, Turkey, iyigun@metu.edu.tr,
Elcin Kartal Koc
3 - Predicting the Effect of Introducing Walk in Hours on Staff
Workload at a Pediatrics Practice
Saurabh Jha, University of Pittsburgh, 1048 Benedum Hall,
Department of Industrial Engineering, Pittsburgh, PA, 15261,
United States of America, saj79@pitt.edu, Louis Luangkesorn,
Diana Hoang, Lindsey Jones, Tricia Pil
A local pediatrics practice has introduced patient walk-in hours in response to
competition from urgent care clinics and has asked to determine the effect on staff
workload. We evaluate the effect of walk-in hours on the practice workload, then
develop and validate a predictive model for the various types of visits and phone
calls. After validating the predictive model, we develop a forecast for the
remainder of the 12 months period following the introduction of all-day walk-ins.
Multivariate Adaptive Regression Splines (MARS) is a popular nonparametric
regression for estimating the nonlinear relationship within data via piecewise
functions. A clustering based knot selection method has been proposed to the
literature recently. This study proposes a parameter selection criteria based on
Schwarz’s Bayesian Information for determining the optimum grid size and
threshold value of this new procedure. Numerical studies are conducted via
artificial and real datasets.
■ TD30
2 - Discovering Interpretable Nonlinear Variation Patterns in
High-Dimensional Data over Spatial Domains
Phillip Howard, Arizona State University, 699 S Mill Ave, Tempe,
AZ, 85281, United States of America, prhoward@asu.edu,
Daniel Apley, George Runger
Decision Support Systems II
Contributed Session
Chair: Rohit Nishant, Assistant Professor, ESC Rennes School of
Business, 2 rue Robert d’Arbrissel CS 76522, Rennes, 35065, France,
rohit.nishant@esc-rennes.com
1 - Routing Recommendation System for Uber
Yuhan Wang, University of California Irvine, 6478 Adobe Cir,
Irvine, CA, 92617, United States of America,
wangyuhan1101@gmail.com
The objective of this research is the identification of distinct and interpretable
nonlinear variation patterns in high-dimensional data through dimensionality
reduction. We present a new method for learning reduced dimension
representations which characterize interpretable variation sources when mapped
to the original feature space. A new metric for measuring how well the solution
can be interpreted is also proposed. We compare our work to alternative methods
using several examples.
3 - EEG Signal Classification using Functional Principal
Component Analysis
Woo-Sik Choi, Mr., Korea University, 145, Anam-Ro, SeongbukGu, Innovation Hall, 816, Korea University, Seoul, Korea,
Republic of, etpist@korea.ac.kr, Seoung Bum Kim
Paper not available at this time.
2 - Optimising Allocation of Investor Funds in Multi-objective Public
Infrastructure Investment Programs
Martin Spollen, Queens University Belfast, David Bates Building,
University Road, Belfast, BT7 1NN, United Kingdom,
mspollen01@qub.ac.uk
Electroencephalogram (EEG) is recordings of the electrical potentials of the brain.
Identifying human activities from EEG is the main goal of brain computer
interface area. To analyze events, selecting important features is a crucial step. In
this study, we propose a feature extraction using functional principal component
analysis with general classification methods. The effectiveness of the proposed
method is demonstrated through a real data from the brain computer interface
competition 2003.
This session will examine the development and application of Strategic
Infrastructure Planning Models (SIPMs) as an emerging class of investment
appraisal techniques for public investment management. The techniques focus
attention on the network effects of investment on total portfolio performance.
Application to a major regional schools investment and rationalization program is
demonstrated.
355
TD32
4 - Modeling Time-varying Autocorrelation for Time
Series Classification
Mustafa Gokce Baydogan, Assistant Professor, Bogaziçi
University, Department of Industrial Engineering, Bebek,
Istanbul, 34342, Turkey, mustafa.baydogan@boun.edu.tr,
George Runger
5 - Enterprise Social Networking and Firm Creativity
Donghyun Kim, Delta State University, 1003 West Sunflower
Road, Cleveland, MS, 38733, United States of America,
dkim@deltastate.edu, Jaemin Kim
This study examines the influence of firm’s IT social networking (SN) capacity on
the firm’s creativity and innovation. Analyzing data on utility patents of 7 firms
using enterprise SN, we tested our predictions on a balanced panel of the firms’
data. The results illustrate how IT SN capacity can aid in the generation of an
idea.
We introduce a novel approach to model the dependency structure in time series
(TS) that generalizes the concept of autoregression to local auto-patterns. A
learning strategy that is fast and insensitive to parameter settings is the basis for
the approach. This unsupervised approach to represent TS generally applies to a
number of data mining tasks. We provide a research direction that breaks from
the linear dependency models to potentially foster other promising nonlinear
approaches.
■ TD33
Decision and Prediction Models in Healthcare
■ TD32
Sponsored Session
Data Mining
Chair: Jakob Kotas, University of Washington, Dept. of Applied
Mathematics, Box 353925, Seattle, WA, 98195,
United States of America, jkotas@uw.edu
1 - A Stochastic Program with Chance-constrained Recourse for
Surgery Scheduling and Rescheduling
Gabriel Zenarosa, PhD Candidate, University of Pittsburgh, 3700
O’Hara Street, Benedum Hall 1048, Pittsburgh, PA,
15261-3048, United States of America, glz5@pitt.edu,
Andrew J. Schaefer, Oleg Prokopyev
Contributed Session
Chair: Gustavo Lujan-Moreno, Arizona State University, Tempe, AZ,
United States of America, glujanmo@asu.edu
1 - Open-source Statistical Packages: The True Cost of
“Free” Software
Ronald Klimberg, Saint Joseph’s University, 35 Moorlinch Blvd.,
Medford, NJ, 08055, United States of America, klimberg@sju.edu,
Rick Pollack, Susan Foltz Boklage
Aggregate surgical expenditures in the US amount to a significant percentage of
GDP. About 42% of hospital revenues are generated by operating rooms (ORs),
yet ORs run at only 68% capacity on average. The most important issues in OR
management are centered on scheduling. Advance schedules improve OR
efficiency; however, surgeries are rescheduled in practice as they rarely go as
planned. We present a stochastic program with chance-constrained recourse for
surgery scheduling and rescheduling.
Open source software is typically free and widely accessible to the public. In the
statistical realm, R is the dominant open-source player. Is it really free? Where do
you go for support? Are their possible significant costs associated with using R?
Further, to what degree should open-source statistical software be used and
taught in academia? This article will explore these questions, as well as others, in
discussing what are often the hidden costs of using open-source statistical
software
2 - Dynamic Scheduling of a Post-discharge Follow-up Organization
to Reduce Readmissions
Sean Yu, Indiana University-Bloomington, 1275 E. Tenth Street,
Bloomington, IN, 47405, United States of America,
xy9@indiana.edu, Shanshan Hu, Jonathan Helm
2 - Study on Effects on Emotional Intensities of Negative Online
Reviews on its Usefulness
Cuiping Li, Huazhong University of Science and Technology, 1037
Luoyu Rd, Hongshan District,Wuhan, Hubei, 430074, China,
412543536@qq.com, Qian Yuan, Shuqin Cai
Hospital readmissions are a growing problem. Many readmissions are preventable
by properly monitoring patients post-discharge. We consider an organization that
dynamically will schedule and staff post-discharge monitoring schedules for a
cohort of patients being randomly discharged from client hospitals. We formulate
this problem as an infinite horizon dynamic program that can be solved using
approximate dynamic programming.
Aimed at recognizing high quality reviews from mass data, this paper explores
how reviews’ negative emotions influence the usefulness of negative online
reviews by using data mining technology and regression analysis. The result
reveals that strong negative emotions reduce negative reviews’ usefulness and
moderate negative emotions have opposite effect. Results also show that different
intensities of negative emotions have significant interactions on reviews’
usefulness.
3 - Predictive Capabilities in Hierarchical Node-based Clustering of
Time Series
Hootan Kamran, PhD Candidate, University of Toronto,
12 Rodney Blvd., North York, ON, M2N4B6, Canada,
hootan@mie.utoronto.ca, Dionne Aleman, Kieran Moore,
Mike Carter
3 - Impact of Library Online Resource use on Students
Academic Outcome
Fan Zhang, University of Pittsburgh, 1048 Benedum Hall,
Department of Industrial Engineering, Pittsburgh, PA, 15261,
United States of America, faz31@pitt.edu, Louis Luangkesorn,
Ziyi Kang, Yunjie Zhang, Shi Tang
Flu activity is shown to be affected by local variables such as climate. Therefore,
localized activity must be monitored to study spatiotemporal spread patterns of
the disease. Using a 10-year flu dataset from 103 hospitals in Ontario, we
compare predictive capabilities extracted from existing aggregation scheme
(LHIN) with those extracted from the novel hierarchical node-based clustering
scheme and show that the new method will extract more statistically significant
predictive capabilities.
University libraries have a need to demonstrate the impact of their resources on
the University mission: academics and research. However, for electronic
resources, research has shown that students often do not recognize they are using
library resources, making surveys and other assessments not useful. We use
undergraduate demographic and academic outcome data along with logs of online
library resource access to determine if relationship exists between online resource
use and academic outcome.
4 - A Stochastic Dynamic Programming Model for
Response-guided Dosing
Jakob Kotas, University of Washington, Dept. of Applied
Mathematics, Box 353925, Seattle, WA, 98195,
United States of America, jkotas@uw.edu, Archis Ghate
4 - A Case-Crossover Study to Evaluate the Effect of Player Affective
State on Performance in Video Game
Gustavo Lujan-Moreno, Arizona State University, Tempe, AZ,
United States of America, glujanmo@asu.edu
We discuss a stochastic dynamic programming (DP) model to assist with dosing
decisions in response-guided dosing (RGD). The goal in this framework is to
deliver the right dose to the right patient at the right time. We present robust,
optimal learning, and optimal stopping variants of this problem. The structure of
optimal policies in these problems will be explored both analytically and
numerically.
Using an electroencephalogram (EEG) headset we examined whether there was a
significant change in the affective state reported by the EEG when a participant
made a mistake while playing a popular video game. There were five affective
constructs that were examined: engagement, frustration, meditation, short and
long term excitement. We propose a case-crossover methodology to analyze this
type of events. Results show that there is a significant difference in three affective
states.
356
■ TD34
TD36
2 - Service-based Distribution Network Model for Location of
Temporary Relief Facilities
Shaligram Pokharel, Professor, Qatar University, Doha, Qatar,
shaligram@qu.edu.qa, Rojee Pradhananga, Fatih Mutlu,
Jose Holguin-Veras
Joint Session HAS/MSOM-Healthcare: Operational
Issues and Information Sharing in Healthcare
A supply allocation and distribution model is proposed to minimize the total
waiting times at the demand points by considering the possibilities of transferring
excess resources between the temporary facilities and backordering of demand in
different time periods. Model is applied on a test instance to analyze the service
and cost trade-offs. (This research is made possible by a NPRP Award NPRP 5-2005-027 from the Qatar National Research Fund (a member of the Qatar
Foundation). The statements herein are solely the responsibility of the authors.)
Sponsored Session
Chair: Subodha Kumar, Carol And G. David Van Houten, Jr. ‘71
Professor, Mays Business School, Texas A&M University, Wehner 301F
- 4217 TAMU, College Station, TX, 77843, United States of America,
skumar@mays.tamu.edu
1 - Sustainability Planning for Healthcare Information Exchanges
Tharanga Rajapakshe, Assistant Professor,
University of Florida, W. University Ave, Gainesville, FL, 32611,
United States of America, tharanga@ufl.edu
3 - The Network Structure: What it Can Tell About Disaster Warning
Effectiveness?
Xiangyang Guan, University of Washington, 201 More Hall,
Box 352700, Seattle, WA, 98195-2700, United States of America,
guanxy@uw.edu, Cynthia Chen
We develop an analytical framework to study sustainability of Healthcare
Information Exchanges and to demonstrate its use when the revenue is generated
(i) under one revenue model (like membership fee), and (ii) under a combination
of multiple revenue models (like membership fee and rebate structure for the
practice from supporting vendors).
Knowing how public awareness and action change after disaster warning is
critical for effectiveness of warning issuance. Multiple data sources – social media,
taxi trips and subway ridership – are leveraged. Temporal evolutions of the
structure (motifs) of social media network and subway network are established as
measures of public awareness and action. Our result identified a lag of one day in
average between warning issuance and public awareness, and between public
awareness and action.
2 - The Impacts of Healthcare Information Exchanges:
An Empirical Investigation
Emre Demirezen, Assistant Professor, Binghamton University
SUNY, 4400 Vestal Parkway East, AA-242, Binghamton, NY,
13902, United States of America, edemirezen@binghamton.edu,
Subodha Kumar, Ramkumar Janakiraman
■ TD36
In the last decade, the U.S. government has been aggressively promoting the use
of electronic health records and the establishment of regional healthcare
information exchanges (HIEs). HIEs facilitate electronic health information
exchange among healthcare providers that is considered to be beneficial for the
society. However, the real benefits of HIEs are not well understood. Hence, in this
study, we work with an HIE provider based in the state of New York to investigate
the benefits of HIEs.
Fire and Emergency Medical Services
Sponsored Session
Chair: Laura Mclay, Associate Professor, University of Wisconsin,
1513 University Ave, ISYE Department, Madison, WI, 53706,
1 - Predicting the Spatial Distribution of Heart Attack Incidence
in Alberta
Armann Ingolfsson, University of Alberta, Edmonton, Canada,
aingolfs@ualberta.ca, Amir Rastpour, Reidar Hagtvedt
3 - Chance Constrained Operating Room Scheduling with Uncertain
and Ambiguous Information
Zheng Zhang, University of Michigan, 1205 Beal Ave, Ann Arbor,
MI, 48105, United States of America, zzhang0409@gmail.com,
Brian Denton, Xiaolan Xie
We describe stochastic programming and distributionally robust optimization
models that allow for uncertain or ambiguous surgery duration data, respectively.
Each of the models considers surgery-to-OR allocation decisions in the context of
probabilistic constraints on completion time that vary by OR. We describe column
generation approaches that are tailored to these two model formulations. Results
are presented to illustrate the potential use of these models in practice.
We use Poisson regression with a linear-without-intercept link function to predict
the incidence of heart attacks by geographic region, as a function of age, gender,
education, and income characteristics. We discuss model specification and
validation and we present results based on 10 years of empirical data.
2 - Dynamic Ambulance Allocation Utilizing Demand Data Analytics
for Pre-hospital EMS
Yu-Ching Lee, Assistant Professor, Department of Industrial
Engineering and Engineering Management, National Tsing Hua
University, No. 101, Section 2, Kuang-Fu Road, Hsinchu, Taiwan
- ROC, yclee@ie.nthu.edu.tw, Albert Chen, Yu-shih Chen
4 - Bundled Payments for Healthcare Services: A Framework for the
Healthcare Provider Selection Problem
Seokjun Youn, PhD Student, Research Asistant, Mays Business
School, Texas A&M University, 320R Wehner Building,, 4217
Texas A&M University, College Station, TX, 77843, United States
of America, syoun@mays.tamu.edu, Chelliah Sriskandarajah,
Subodha Kumar
Pre-hospital Emergency Medical Services (EMS) provide the critical function of
on-site medical treatment and stabilization of patients. The quality of EMS affects
the survival of patients in emergency situations. A better management of
ambulances could potentially improve the effectiveness and efficiency of EMS.
We study a real-time decision support system featuring demand prediction,
distribution estimation, scenario generation, robust ambulance deployment, and
the optimal ambulance dispatching
Identifying competitive healthcare providers is an important issue for the
successful operation of bundled payments. We develop a selection framework via
data envelopment analysis and combinatorial auction (CA). Based on efficiency
and effectiveness measures, outstanding performers are pre-selected. Finally, CA
determines winners. To evaluate the impact of design issues on the CA
performance, we combine and utilize several real dataset from the healthcare
sector.
3 - A Simulation Optimization Method for Emergency
Service Location
Rozhin Doroudi, PhD Student, Northeastern University, 260
Washington St. Apt. 305, Malden, MA, 02148, United States of
America, doroudi.r@husky.neu.edu, Gerald Evans, Gail Depuy
■ TD35
A fire department with an available fleet wants to determine the location of its
fire stations and how to distribute its fleet among those locations. An iterative
approach involving the use of a linear program and a simulation model is
proposed for the problem
Disaster and Emergency Management II
Contributed Session
Chair: Shaligram Pokharel, Professor, Qatar University, Doha, Qatar,
shaligram@qu.edu.qa
1 - Solution Methodologies for Debris Removal in Disaster Response
Bahar Y. Kara, Associate Professor, Bilkent University, Industrial
Engineering, Ankara, Turkey, bkara@bilkent.edu.tr,
Oya E. Karasan, Nihal Berktas
In this study we provide solution methodologies for debris removal problem in
the response phase. Debris removal activities on certain blocked arcs have to be
scheduled in order to reach a set of critical nodes such as schools and hospitals.
Two mathematical models are developed with different objectives. The models are
tested over real data from districts of Istanbul.
357
TD37
■ TD37
2 - Moments of a Random Set
Kemal Gursoy, Rutgers University, 100 Rockafeller Road,
Department of MSIS, Piscataway, NJ, 08854,
United States of America, kgursoy@rci.rutgers.edu
Therapy and Treatment
Let X be a random subset of the n-dimensional Euclidean space. Then the
moments of the measure of X could be constructed by the Lebesgue integral of
the probability measure of every point in X, over the n-dimensional Euclidean
space.
Contributed Session
Chair: Animesh Garg, PhD Student, UC Berkeley, 4141 Etcheverry Hall,
Berkeley, CA, 94720-1777, United States of America,
garg.animesh@gmail.com
1 - Assessment of a Novel Device for Elbow Rehabilitation in Humans
Aline Callegaro, Researcher, UFRGS, 99 Osvaldo Aranha Avenue,
5 Floor, Porto Alegre, RS, 90035190, Brazil, nimacall@gmail.com,
Carlos Fernando Jung, Clarissa Brusco,
Marcelo Gava Pompermayer, Márcia Elisa Echeveste,
Carla Schwengber ten Caten
3 - A New Representation for the Stationary Distribution of
Markov Chains
Patrick Buckingham, Clemson University, Mathematical Sciences,
Clemson, SC, United States of America, pbuckin@clemson.edu,
Brian Fralix
We present a new representation for the stationary distribution of ergodic Markov
chains, as well as analogous representations for Laplace transforms of transition
functions associated with such chains. Applications to hysteretic queues and other
models will be discussed.
This study aimed to assess a novel device for elbow rehabilitation in humans. The
functional prototype assessment was based on data collection in two stages:
application of local muscle vibration; and its association with Continuous Passive
Motion. Two way ANOVA was used to analyse the main factors. An average
increase of the muscle electrical activation resulted in first stage. The main factors
(frequency, sex, and treatment) had significant effect, as well as a few
interactions.
4 - Queueing Systems with Adaptive Service Rates
Raik Stolletz, University of Mannheim, Room SO 230,
Schloss Schneckenhof Ost, Mannheim, Germany,
stolletz@bwl.uni-mannheim.de, Jannik Vogel
In many service systems, for example call centers, the service rate could be
considered as a time-dependent decision variable to improve and stabilize the
performance of a queueing system. We present an M(t)/M(t)/c queueing model
with adaptive service rates. The SBC-approach is used to approximate the timedependent behaviour of by stationary models. This results in non-linear
optimization problem. Numerical examples show the benefits of assuming the
service rate as a decision variable.
2 - Minimizing Metastatic Risk in Radiotherapy
Fractionation Schedules
Hamidreza Badri, Graduate Student, University of Minnesota,
ISyE Departemnt, 111 Church Street S.E., Minneapolis, MN,
55455, United States of America, badri019@umn.edu,
Jagdish Ramakrishnan, Kevin Leder
5 - Modeling with the Tilted Beta Distribution
Gene Hahn, Associate Professor, Salisbury University, 1101
Camden Ave., Salisbury, MD, 21801, United States of America,
edhahn@salisbury.edu
The treatment of metastatic cancer remains an extremely challenging problem.
Here we consider the problem of developing fractionated irradiation schedules
that minimize production of metastatic cancer cells. We observe that the resulting
fractionation schedules are different than those that result from more standard
objectives such as minimization of final primary tumor volume. Delivering large
doses in small fractions is suggested even in cases when a/b value of the tumor is
large.
The beta distribution has an important limitation for the modeling of bounded
data. Its density is either zero or infinite at the endpoints except for special cases.
This makes modeling certain kinds of data difficult. The tilted beta distribution can
be used to easily model this data. We adopt a Bayesian perspective and examine
its usage in modeling real-world data.
3 - Customized 3D Printed Implants with Internal Channels for
Intracavitary High Dose Rate Brachytherapy
Animesh Garg, PhD Student, UC Berkeley, 4141 Etcheverry Hall,
University of California, Berkeley, CA, 94720-1777,
United States of America, garg.animesh@gmail.com, Jean Pouliot,
J. Adam M. Cunha, I-Chow Hsu, Alper Atamturk, Ken Goldberg
■ TD39
39-Room 100, CC
High-Dose Rate Brachytherapy is an internal radiation therapy frequently used
for cancer treatment. Radioactive sources are briefly placed proximal to tumors.
Current methods for intracavitary HDR-BT use generic templates which limits
dose distribution to a small set of linear channels. We propose the use of
algorithmically customized 3D Printed implants with curved internal channels
that fit cavities without tissue puncture and aim to improve dose distribution and
treatment quality.
Product Brand Differentiation and Pricing Decisions
Invited Session
Chair: Ruixia Shi, California State University, Fullerton, 800 N. State
College Blvd., Fullerton, CA, 92834, United States of America,
sandy.shi@gmail.com
1 - New Product Pricing Strategy in the Social Media Era
Gou Qinglong, Associate Professor, University of Science &
Technology of China, No.96, JinZhai Road Baohe District, Hefei,
China, tslg@ustc.edu.cn, Kumar Subodha, Xiuli He, Juzhi Zhang
4 - Surgery Sequencing and Scheduling in Multiple ORs with
PACU Constraints
Miao Bai, Lehigh University, 200 W Packer Ave, Bethlehem, PA,
18015, United States of America, mib411@lehigh.edu,
Gregory Tonkay, Robert Storer
We study a multiple-OR surgery sequencing and scheduling problem with PACU
constraints. To minimize the cost incurred by waiting, idleness, OR blocking and
overtime, a two-stage solution scheme is proposed. In the first stage, a timeindexed integer program is formulated and solved by Lagrangian relaxation and
dynamic programming to determine surgery sequences. Given surgery sequences
in all ORs, scheduled times of patients are found in the second stage by a samplegradient-based algorithm.
With the popularity of various social media platforms, the impacts of the word of
mouth effect and the reference price effect on a consumer’s purchasing behavior
have been significantly amplified in the current era. We incorporate these two
effects into a two period pricing model to investigate whether and in which
condition should a firm utilize a skimming or a penetration price strategy. Our
results show how these two effects influence a firm’s pricing strategy when he
launches a new product.
2 - Competition and Coordination in Online Retailers and
Express Companies
Yihong Hu, Assistant Professor, Tongji University,
yhhu@tongji.edu.cn
■ TD38
Probability
We consider an online market as Taobao with homogenous and heterogeneous
consumers sensitive to service quality. Retailers collect a separate product price
plus shipping fee from consumers. They treat shipping fee as a source of revenue
by asking a large discount of shipping fee from shippers. We use game-theoretic
framework to study the competition and coordination between retailers and
shippers under different scenarios. The study finds that retailers’ behavior
increases consumers’ benefit.
Contributed Session
Chair: Gene Hahn, Associate Professor, Salisbury University,
1101 Camden Ave., Salisbury, MD, 21801, United States of America,
edhahn@salisbury.edu
1 - Fluctuation Analysis and the Marked Poisson Process
Randy Robinson, Bemidji State University, Bemidji, MN,
United States of America, rrobinson@bemidjistate.edu
This presentation studies the marked point process with position dependent
marking. The focus was on predicting the first passage time of the marked
random walk when exiting a given rectangular set and the value of the process
upon the exit. A new density function for the related processes has been
obtained: a product of a negative exponential and modified Bessel functions
358
TD41
3 - Supply Chain Power and Store Brand
Jun Ru, Assistant Professor, University at Buffalo, 326D Jacobs,
Buffalo, NY, 14260, United States of America, junru@buffalo.edu,
Ruixia Shi, Jun Zhang
5 - An Empirical Study to Examine Consumer Behavior towards
Luxury Brands in Pakistan
Faryal Salman, Assistant Professor, SZABIST, 90 Clifton, Karachi,
Karachi, Pakistan, faryal.salman@szabist.edu.pk, Usman Warraich
This paper relates a retailer’s store brand strategy to the relative powers of
channel members and offers a new explanation for the differences in retailers’
store brand strategies. Our analysis shows that store brands become less appealing
to a retailer as it becomes more powerful.
Current study seeks to expand an understanding of consumer behavior towards
branded goods. The data for this exploratory study was collected from urban
youth of Pakistan. The study postulates significant relationship between consumer
behavior and the predictors for various product categories. Regression analysis
shows that these variables pose the positive impact on the buying behavior (p
value (0.05) and this model shows R2 of 0.73.
4 - Consumer Preference Mismatch and Channel Choice Decisions
under Competition
Kunpeng Li, Utah State University, 3555 Old Main Hill,
Logan, UT, United States of America, kunpeng.li@usu.edu,
Suman Mallik, Dilip Chhajed
■ TD41
We consider a product consisting of two components sold by two firms. A
product/firm is integrated when both components are designed by a single firm,
and is non-integrated otherwise. The consumers choose to purchase a product
that better matches the specifications of their ideal product. Using a duopoly
model, we study the effects of consumer preference mismatch on channel
integration strategies.
41-Room 102A, CC
Healthcare Supply Chain Decision Making
Sponsored Session
Chair: Xinghao Yan, Assistant Professor, Ivey Business School, Western
University, 1255 Western Road, London, On, N6G0N1, Canada,
xyan@ivey.uwo.ca
1 - Determinants of Distribution Channel Choice in Pharmaceutical
Industry – Specialty Drugs
Liang (Leon) Xu, University of Missouri-St. Louis, St. Louis, MO,
United States of America, lxpx2@umsl.edu, Vidya Mani, Hui Zhao
■ TD40
40- Room 101, CC
Marketing II
Contributed Session
Chair: Faryal Salman, Assistant Professor, SZABIST, 90 Clifton,
Karachi, Pakistan, faryal.salman@szabist.edu.pk
1 - An Analysis of Menus of Multi-Part Tariffs
Ryan Choi, PhD Candidate, UC Irvine, 6219 Adobe Circle, Irvine,
CA, 92617, United States of America, jihungc@uci.edu
We use privately collected multi-year transaction data to study determinants of
the choice of distribution channels for specialty and non-specialty drugs. Further,
we explore how this channel choice explains observed variations in supply chain
metrics in this industry.
2 - Transforming Drug Development via System
Computational Modeling
Jinha Lee, Georgia Institute of Technology, 755 Ferst Dr. NW,
Atlanta, GA, United States of America, jlee68@gatech.edu,
Eva Lee
This paper study which characteristics of three-part tariffs make the seller more
profitable than two-part tariffs. Given a full extraction of low type segment’s
surplus, the seller can extract more of high type surpluses, whose magnitude is
dependent on both of the level of quantity allowances and the fixed fee for high
type consumers. With 3PTs, firms earn more rent from the high type, and so
offers both high and low contracts regardless of the taste parameter and of the
low type proportion.
We describe the firstin-silico drug design system model to accelerate drug
discovery. The model spans preclinical, clinical, IND and NDA tasks; and allows
global risk analysis. It identifies bottlenecks, and performs system optimization
that offers a holistic view of discovery pathways. Rapid development is achieved
through parallel processes that shorten critical paths from start to registration of a
new drug. The generalizable design allows rapid testing, and minimizes risk, cost,
and time.
2 - Research and Practice – Friends of Foes? Perceptions of
Marketing Academicians and Practitioners.
Salma Rahman, Assistant Professor, SZABIST, 100 Clifton,
Block 5, Shahrae Iran, Karachi, Pakistan, sal_haider@yahoo.com,
Sana Rehman
3 - Operational Performance Evaluation of Reverse Referral
Partnership in the Chinese Healthcare System
Nan Kong, Associate Professor, Purdue University, 206 S. Martin
Jischke Dr., West Lafayette, IN, United States of America,
nkong@purdue.edu, Quanlin Li, Na Li, Zhibin Jiang
This research is more of an exploratory nature that focuses initially on
precipitating the perceptions of marketing academicians as well as marketing
practitioners regarding the existence of the research practice gap using diffusion
of innovations theory. The results indicated that generally both agree on the
prevalence of the gap. Further, their perception is the same for discovery and
translations stage whereas it differs for the dissemination and change stage.
Reverse referral of patients from upper-level hospitals to lower-level hospitals
after their acute care, has been promoted in the tiered Chinese care system to
alleviate resource pressure at high-level hospitals and balance utilizations
throughout the system. However, it remains unclear how to implement reverse
referral partnerships given the conflicting interests. We develop a two-level
queuing network model to capture patient flows and derive analytical results on
queueing performance measures. Our work is expected to guide the
establishment of hospital alliances in China.
3 - Customer Commitment in Customer Churn Prediction
Huili Liu, Beijing University of Posts and Telecommunications,
Xitucheng Road, No 10, Beijing, BJ, 100876, China,
yucailhl@163.com
We present a customer commitment model to predict the customer churn. Instead
of probability method, we use consumer learning to get the customer
commitment from his/her purchase history, and then use it to predict the
customer churn. In comparison to existing models, we consider the customer
commitment model is more feasible with an accurate prediction. Thereby we
provide a new insight into the customer base analysis.
4 - Influenza Vaccine Supply Chain with Vaccination Promotion Effort
and its Coordination
Xinghao Yan, Assistant Professor, Ivey Business School, Western
University, 1255 Western Road, London, ON, N6G0N1, Canada,
xyan@ivey.uwo.ca, Gregory Zaric
4 - Socio-economic Class Difference in Movie Consumption
Among Pre-adolescents
Saima Husain, Lecturer, Institute of Business Administration,
University Campus, University Road, Karachi, Si, Pakistan,
shusain@iba.edu.pk
We develop an influenza vaccine supply chain model consisting of a health
authority, a vaccine manufacturer, and population. The health authority decides
order quantity and effort exerted to increase vaccination demand; the
manufacturer decides production effort; and population decides the vaccination
probability. We find that the three parties’ decisions at equilibrium and different
coordinating contract formats, such as a contract with payment linear/piecewise
linear w. r. t. order quantity.
This research uses the laddering technique, in semi-structured in depth
interviews, to study hierarchical constructs explaining personal value system that
drive movie consumption behaviour among children aged 9 - 12 years. Young
informants were recruited from different socio economic class (SEC) households
in Pakistan. Findings show that children from higher SECs are significantly
different in the type of movie selection, consumption setting and medium used
for movie consumption.
359
TD42
■ TD42
3 - Optimal Pricing for a Multinomial Logit Choice Model with
Network Effects
Chenhao Du, Student, University of Minnesota, 425 13th Ave SE,
Apt. 1502, Minneapolis, MN, 55414, United States of America,
duxxx181@umn.edu, William Cooper, Zizhuo Wang
42-Room 102B, CC
Patients and Practice: Using the Right Resources to
Deliver Care
We consider a seller’s problem of determining revenue-maximizing prices for an
assortment of products that exhibit network effects. Customers make purchase
decisions according to a modified MNL choice model. We show that the optimal
strategy is either to maintain a semblance of balanced sales among all product or
to boost the sales of exactly one product. We also show the importance of taking
the network effects into consideration.
Sponsored Session
Chair: Jonathan Helm, Indiana University Bloomington, 1309 E. Tenth
Street, Bloomington, IN, United States of America, helmj@indiana.edu
1 - An Empirical Study of The Impact of Physician Assistants During
Critical Care Consultations
Yunchao Xu, New York University, 44W 4th St, 8-152, New York,
NY, 10012, United States of America, yxu4@stern.nyu.edu,
Carri Chan, Mor Armony
4 - Pricing Ancillary Service Subscriptions
Ruxian Wang, Johns Hopkins University, 100 International Dr,
Baltimore, MD, 21202, United States of America,
ruxian.wang@jhu.edu, Maqbool Dada, Ozge Sahin
We investigate customer choice behavior in the presence of main products,
ancillary services with options of pay-per-use and subscription, as well as the
outside option. Analytical results and numerical experiments show that offering
service subscriptions may result in “win-win-win”“win-win-lose”“lose-lose-win”
and other situations for the firm, competitors and customers in the monopolistic
and competitive scenarios.
Trained with a broad set of clinical skills, physician assistants (PAs) can be costeffective alternatives to physicians in healthcare systems. However, not much is
known on the impact of PAs on patient delivery in certain settings. Using data
from a major urban hospital system, we utilize a difference-in-differences
approach to explore the effects of introducing PAs into the critical care
consultation process. One key finding is the reduction in boarding times due to
this intervention.
■ TD44
2 - Missed Opportunities in Preventing Hospital Readmissions:
Redesigning Post-discharge Checkup Policie
Xiang Liu, University of Michigan, 1205 Beal Ave, Ann Arbor,
MI, 48109, United States of America, liuxiang@umich.edu,
Jonathan Helm, Ted Skolarus, Michael Hu, Mariel Lavieri
44-Room 103B, CC
Recent Trends in Retailing
Sponsored Session
Hospital readmissions affect hundreds of thousands of patients, placing a
tremendous burden on the healthcare system. Post-discharge checkup can reduce
readmissions through early detection of conditions. Our work develops optimal
checkup plans to monitor patients following hospital discharge using methods
including phone calls and office visits. By analyzing the structure of optimal
policies, we develop checkup schedules that mitigate 32% more readmissions.
Chair: Mehmet Sekip Altug, Assistant Professor, George Washington
University, Washington, DC, United States of America,
maltug@gwu.edu
1 - Analyzing Big-Box Retailer in an Emerging Market
Mehmet Gumus, McGill University, 1001 Sherbrooke Street West,
Montreal, Canada, mehmet.gumus@mcgill.ca, Aditya Jain,
Saibal Ray
3 - Incentive-compatible Prehospital Triage in Emergency
Medical Services
Eric Webb, Graduate Student, Indiana University, 1309 E. 10th
Street, Bloomington, IN, 47405, United States of America,
ermwebb@indiana.edu, Alex Mills
We consider the impact of the entry of a big-box retailer in a market dominated
by small retailers. The small retailers are characterized by local coverage of the
market, whereas the big-box retailer provides services valued by customers. Since
both types of retailers obtain supplies from a common manufacturer, big-box
retailer’s entry affects the supply conditions. Our work thus highlights roles of
direct competition as well as indirect supply side effect on small retailers and
customers.
The Emergency Medical Services (EMS) system is designed to handle lifethreatening emergencies, but a large and growing number of non-emergency
patients seek healthcare through EMS. We evaluate the incentives underlying
prehospital triage, where EMS staff are allowed to identify patients that could be
safely diverted away from the hospital and toward appropriate care. Continued
transition from fee-for-service payments to bundled payments may be necessary
for prehospital triage implementation.
2 - Dynamic Pricing with Customer Upgrades
Oben Ceryan, Assistant Professor, Drexel University, 3220 Market
St., Philadelphia, PA, United States of America, oc43@drexel.edu,
Ozge Sahin, Izak Duenyas
We study the impact of product upgrades on a firm’s pricing and replenishment
policies by considering a multiple period, two-stage model where the firm first
sets prices and replenishment levels, and after observing the demand, it decides
whether to upgrade any customers to a higher quality product. We characterize
the structure of the optimal upgrade, pricing, and replenishment policies and find
that offering upgrades assists in preserving the vertical price differentiation of the
products.
■ TD43
43-Room 103A, CC
Revenue Management with Consumer
Choice Models
Sponsored Session
3 - Return Abuse, Countermeasures, and Privacy Concerns
Serkan M. Akturk, PhD Candidate, Texas A&M University, 4217
TAMU Wehner 320 M, College Station, TX, United States of
America, makturk@mays.tamu.edu, Michael Ketzenberg
Chair: Ruxian Wang, Johns Hopkins University, 100 International Dr,
Baltimore, MD, 21202, United States of America, ruxian.wang@jhu.edu
1 - Dynamic Pricing for Mobile Apps
Kejia Hu, Kellogg School of Management, Northwestern
University, 2169 Campus Drive, Evanston, IL, United States of
America, k-hu@kellogg.northwestern.edu, Chaitanya Bandi,
Srikanth Jagabathula
This paper analytically investigates return abuse with respect to both fraudulent
and opportunistic consumer returns and potential countermeasures to deal with
them. The research also shows how those countermeasures impact a retailer’s
profitability, demand structure, and policy parameters with respect to price and
refund. To some extent, our findings contradict common suggestions in the
literature.
Mobile apps is special in the following aspects. It has no inventory constraint,
almost zero marginal cost and free version updates. In our research, we will
model these features and show the dynamic pricing for mobile apps.
4 - Store-clearance or Secondary Markets? Evaluation of Inventory
Clearance Opportunities in Retailing
Mehmet Sekip Altug, Assistant Professor, George Washington
University, Washington, DC, United States of America,
maltug@gwu.edu, Garrett Van Ryzin
2 - Product Line Design and Pricing under Logit Model
Anran Li, Columbia University, 345 Mudd, New York, NY, 10027,
United States of America, al2942@columbia.edu,
Guillermo Gallego, Jose Beltran
One main assumption in the newsvendor model is that the salvage value is
exogenous and retailers can sell their excess stock at this fixed salvage value.
However, the salvage value of excess stock is mostly determined endogenously.
We compare consolidated secondary markets vs. store clearance with myopic and
strategic customers.
We study a firm who wants to design and price a set of products characterized by
a number of features where each feature has one or multiple levels. We model
consumers’ demand by a feature-level based Logit model and optimize the
assortment on the features space. We find a price independent index of each
feature level that plays a key role. This makes a greedy algorithm, derived from
the K-shortest paths algorithm, able to find an optimal K products’ configuration
in polynomial time.
360
■ TD45
TD47
monotonic-neither most nor least connected customers are prime targets for
making the product available.
45-Room 103C, CC
3 - Supply Disruptions and Optimal Network Structures
Kostas Bimpikis, Stanford GSB, 655 Knight Way, Stanford, CA,
94305, United States of America, kostasb@stanford.edu,
Ozan Candogan, Shayan Ehsani
Topics in Dynamic Pricing and Revenue Management
Sponsored Session
This paper studies multi-tier supply chain networks in the presence of disruption
risk. Firms compete with one another by participating in one of K production
stages. We provide a characterization of the equilibrium prices, profits, and
sourcing decisions and derive insights on how the network structure and the
reliability of production in different tiers affect firms’ profits and the prices of
intermediate goods.
Chair: Robert Phillips, Columbia Business School, 2790 Broadway Uris
Hall, New York, NY, 10027, United States of America,
rp2051@columbia.edu
1 - Dynamic Pricing with Demand Covariates
Sheng Qiang, Student, Stanford University, 41 Olmsted Road,
Apt 108, Stanford, CA, 94305, United States of America,
sqiang@stanford.edu, Mohsen Bayati, Michael Harrison
4 - Creating Reciprocal Value through Operational Transparency
Ryan Buell, Harvard Business School, Morgan Hall 429, Boston,
MA, 02163, United States of America, rbuell@hbs.edu, Tami Kim,
Chia-Jung Tsay
A firm sells products over T periods, without knowing the demand function. The
firm sets prices to earn revenue and learn the demand function. In each period
before setting the prices, the firm observes some demand covariates, like
marketing expenditure, consumer’s attributes, etc. The performance is measured
by the regret, which is the expected revenue deviation from the optimal pricing
policy when demand function is known. We study the asymptotic near-optimal
algorithms to optimize the regret.
We investigate whether organizations can create value by introducing visual
transparency between consumers and producers. Two field and three laboratory
experiments in food service settings suggest that transparency that 1) allows
customers to observe operational processes and 2) allows employees to observe
customers not only improves customer perceptions, but also increases service
quality and efficiency.
2 - What Really Happens in Implementing Revenue Management
Capabilities and What to Expect in the Future
Vedat Akgun, Director, Revenue Analytics, 3100 Cumberland
Blvd SE, Suite 1000, Atlanta, GA, 30339, United States of
America, vakgun@revenueanalytics.com, Jon Higbie
■ TD47
47-Room 104B, CC
Implementation of Revenue Management started more than thirty years ago and
Revenue management concepts have evolved over time providing extraordinary
benefits to companies. In addition to realizing great success, we also face
challenges and learn lessons based on our experience and research. We want to
discuss what really happens in implementing Revenue Management capabilities
and what we can expect in the future.
Sustainable Operations Management
3 - Nonparametric Algorithm for Joint Pricing and Inventory Control
with Lost-sales and Censored Demand
Boxiao (Beryl) Chen, University of Michigan-Ann Arbor, 1205
boxchen@umich.edu, Xiuli Chao, Cong Shi
Chair: David Drake, Assistant Professor, Harvard Business School,
Morgan Hall 425, Boston, MA, United States of America,
ddrake@hbs.edu
1 - Mobile Money Agent Inventory Management
Karthik Balasubramanian, Harvard Business School, 25 Harvard
Way, Boston, MA, 02163-1011, United States of America,
kbalasubramanian@hbs.edu, David Drake, Douglas Fearing
Operations
Sponsored Session
We consider the classic joint pricing and inventory control problem with lost-sales
and censored demand in which the demand distribution is not known to the firm
a priori. Conventional learning algorithms are not applicable as the firm can
observe neither the realized value nor any derivative information of the true
objective function, and the estimate of the expected profit function from data is
not unimodal. We develop a data-driven algorithm which converges and provide
its convergence rate.
Mobile money agents exchange cash for electronic value and vice versa, forming
the backbone of an emerging electronic currency ecosystem in the developing
world. Unfortunately, low agent service levels are a major impediment to the
further development of these ecosystems. We model the agent’s inventory
problem and numerically determine optimal quantities. Finally, we evaluate our
recommendations with a large dataset of mobile money agent transactions in an
East African country.
■ TD46
2 - Energy Efficiency Contracting in Supply Chains under Asymmetric
Bargaining Power
Ali Shantia, HEC-Paris, 7, Avenue De La Gare, Bievres, 91570,
France, ali.shantia@hec.edu, Andrea Masini
46-Room 104A, CC
Service Operations
Mgmt/Service Operations
Sponsored Session
In a supply chain, consisting of a buyer and a supplier, this study analyzes the
effect of relative bargaining power and technology uncertainty on the supplier’s
decision to invest in energy efficiency (EE) measures. We analyze price
commitment and shared investment contracts and compare the two mechanisms
in their ability to boost EE investment when the buyer’s high bargaining power in
addition to high technology uncertainty prevent the supplier from investing in
EE.
Chair: Gad Allon, Professor, Kellogg School of Management,
Northwestern University, 2001 Sheridan Road, Evanston, IL, 60201,
United States of America, g-allon@kellogg.northwestern.edu
1 - Managing Service Systems in Presence of Social Networks
Gad Allon, Professor, Kellogg School of Management,
Northwestern University, 2001 Sheridan Road, Evanston, IL,
60201, United States of America,
g-allon@kellogg.northwestern.edu, Dennis Zhang
3 - Competitive Industry’s Response to Environmental Tax Incentives
for Green Technology Adoption
Anton Ovchinnikov, Queen’s University, 143 Union Str West,
Kingston, Canada, anton.ovchinnikov@queensu.ca, Dmitry Krass
We consider operational aspects of how an industry composed of heterogeneous
firms responds to an environmental tax by choosing production quantities and
emissions-reducing technologies. We show the existence and uniqueness of the
“market-only equilibrium” and demonstrate its many interesting properties. We
then discuss the technology-and-market equilibria under different structural
assumptions.
We study a service system with the presence of a social network. In our model,
firms can differentiate resource allocations among customers, and customers learn
the service qualities from the social network. We study the interplay among
network structure, customer characteristics, and information structure, and
characterize the optimal policy. We further calibrate our model with data from
Yelp.com and quantify the value of social network knowledge empirically.
4 - Carbon Tariffs: Effects in Settings with Technology Choice and
Foreign Production Cost Advantage
David Drake, Assistant Professor, Harvard Business School,
Morgan Hall 425, Boston, MA, United States of America,
ddrake@hbs.edu
2 - Keeping Up with the Joneses: using Social Network Information
to Manage Availability
Ruslan Momot, INSEAD, Boulevard de Constance,
Fontainebleau, 77305, France, ruslan.momot@insead.edu,
Elena Belavina, Karan Girotra
When firms can choose from a set of potential production technologies and
offshore facilities hold a production cost advantage, I show that carbon leakage
due to offshoring and/or foreign entry can result despite the implementation of a
carbon tariff. However, in such a setting, carbon leakage is shown to conditionally
decrease global emissions, contradicting prevailing popular opinion and widely
reported results that do not account for technology choice or foreign production
cost advantage.
Growing availability of data on the patterns of customers’ social interactions has
opened up new opportunities for businesses. We identify an optimal distribution
strategy for a firm selling to socially connected customers engaged in social
comparison. We build a stylized game-theoretic model of strategically interacting
customers in a general network. We find that the optimal strategy is non
361
TD48
■ TD48
2 - Competition and Perceptions of User Reviews
Michael Galbreth, Associate Professor Of Management Science,
Moore School of Business, University of South Carolina,
Columbia, SC, United States of America, galbreth@moore.sc.edu,
Pelin Pekgun, Bikram Ghosh
48-Room 105A, CC
New Directions at the Interface of Finance,
Operations, and Risk Management
We analyze the interaction of user reviews and valuation uncertainty for
experience goods, with a specific focus on the potential for negative vs. positive
reviews to be weighted differently by consumers. The competitive impact of this
unequal weighting is not always intuitive. For example, we show that if a lower
quality firm has a large user base, overweighting of negative reviews can lead to
higher profits and higher prices in equilibrium than its higher quality competitor.
Sponsored Session
Chair: Gerry Tsoukalas, Assistant Professor, Wharton, 3730 Walnut
street, Philadelphia, PA, 19104, United States of America,
gtsouk@wharton.upenn.edu
3 - Analysis of Consumers’ Purchase Timing Decisions
Emre Ertan, PhD Candidate, UT Dallas, Sm30 Jindal School of
Management, 800W Campbell Dr, Richardson, TX, 75080,
United States of America, emre.ertan@utdallas.edu, Kathy Stecke,
Ozalp Ozer
Co-Chair: Vlad Babich, Georgetown University, Washington, D.C.,
Volodymyr.Babich@georgetown.edu
1 - Supply Chain Contract Design under Financial Constraints and
Bankruptcy Costs
Panos Kouvelis, Professor, Olin Business School, Washington
University in St. Louis, St. Louis, MO, United States of America,
kouvelis@wustl.edu, Wenhui Zhao
The consumer purchase timing decision is analyzed by using discounted expected
utility theory, where consumers act to maximize their utility over time. The
consumer’s sequential decision-making process is formalized under uncertain
product availability. An optimal purchase timing policy is identified in a market
environment, in which a strategic customer knows the markdown pricing
scheme, available inventory level, and remaining time to the end of the selling
horizon.
We study contract design in a supply chain of two capital constrained firms in
need of short-term financing. The failure of loan repayment leads to bankruptcy
with fixed and variable default costs. With only variable default costs, buyback
contracts remain equivalent to revenue-sharing contracts, which coordinate with
working capital adjustments. With fixed default costs, a revenue-sharing contract
with working capital coordination might have higher expected profit than the
one-firm system.
4 - Product Line Design and Capacity Management:
The Role of Consumer Behavior Uncertainty
Muge Yayla-Kullu, RPI, 110 8th St, Troy, NY, 12180,
United States of America, Yaylah@rpi.edu, Jennifer Ryan,
Jayashankar Swaminathan
2 - Network Recovery using Transactional Information
John Birge, Professor, University of Chicago Booth School of
Business, 5807 S Woodlawn Ave, Chicago, IL, 60637,
United States of America, john.birge@chicagobooth.edu
We study the effects of uncertainty in consumer spending due to economic
volatility on the product line decisions of a firm with limited resources. We
consider a firm that offers products with differing qualities, unit production costs,
and resource consumption rates. Making capacity allocation decisions in the face
of such an uncertainty is challenging, demanding careful consideration of product
variety and available resources.
Firms operate as components of complex networks of physical and financial
flows. The structure of these networks is however not easily observed. This talk
will discuss methodologies to uncover such hidden structure using inverse
optimization techniques.
5 - A Manufacturer’s Outlet Decision: The Impact of Quality,
Innovation and Market Awareness
Jennifer Ryan, RPI, ISE, CII, Troy, NY, 12180,
United States of America, ryanj6@rpi.edu, Daewon Sun
3 - Does Operational Investment Vary with Capital Structure?
Vishal Gaur, Cornell University, 321 Sage Hall, Ithaca, NY, 14850,
United States of America, vg77@cornell.edu, Yasin Alan
We investigate the relationship between the operational investment of firms and
their capital structure choices using data for U.S. manufacturing and retail trade
sectors.
We consider a manufacturer of a luxury good who must determine whether to
sell products only through a manufacturer-owned retail store, or to also sell
products through the factory outlet store. We study how this decision depends on
the relative qualities of the products offered on the two channels, as well as the
manufacturer’s ability to innovate and introduce new product lines. In addition,
our multi-period model captures the impact of market share on the
manufacturer’s brand awareness.
4 - Entrepreneurial Finance: Crowdfunding, Venture Capital, and
Bank Financing
Vlad Babich, Georgetown University, Washington, DC,
United States of America, Volodymyr.Babich@georgetown.edu,
Gerry Tsoukalas
We study the interplay between bank financing, venture capital and
crowdfunding, in a multi-stage bargaining game, with double-sided moral hazard.
We find that while crowdfunding usually serves a positive role, enabling funding
for good projects, and avoiding investments in bad projects, it may also hurt VCs,
entrepreneur, and the society.
■ TD50
50-Room 106A, CC
Supply Network Management: Collaboration
and Competition
Sponsored Session
■ TD49
49-Room 105B, CC
Chair: Hyoduk Shin, UC-San Diego, San Diego, CA,
United States of America, hshin@rady.ucsd.edu
1 - Optimal Procurement in Assembly Supply Chains:
Contracting Timing and Supplier Mergers
Bin Hu, Assistant Professor, UNC Kenan-Flagler Business School,
CB#3490 McColl Bldg, University of North Carolina, Chapel Hill,
NC, 27519, United States of America, Bin_Hu@kenanflagler.unc.edu, Anyan Qi
Demand Driven Supply Chains
Sponsored Session
Chair: Muge Yayla-Kullu, RPI, 110 8th St, Troy, NY, 12180,
United States of America, YAYLAH@rpi.edu
1 - The Effect of Targeted Coupons on Product Quality Assortment
and Competition
Amit Eynan, Professor, University of Richmond, 1 Gateway Rd,
Richmond, VA, 23173, United States of America,
aeynan@richmond.edu, Benny Mantin
OEMs often procure components from several suppliers to assemble into
products. Such an OEM needs proportional component quantities, calling for a
coordinated procurement mechanism. We propose the use of two-part tariff
contracts for coordinated procurement. We further show that simultaneous and
sequential contracting are equivalent. Finally, we investigate the impact of a
supplier merger in an assembly supply chain.
Manufacturers who sell to customers with heterogeneous valuation of quality can
segment the market by offering multiple products at different qualities and prices.
We investigate the effect of targeted marketing efforts (coupons) on product line
assortment of a monopolist as well as under competition. While coupons help the
monopolist, in the competitive setting, we find that both firms end up exerting
marketing efforts but only one of them is better off whereas the other is worse
off.
362
TD52
2 - Long-Term Partnership for Achieving Efficient Capacity Allocation
Fang Liu, Assistant Professor, Nanyang Business School, Nanyang
Technological University, 50 Nanyang Avenue, South Spine
S3-B2A-13, Singapore, 639798, Singapore, liu_fang@ntu.edu.sg,
Tracy Lewis, Nataliya Kuribko, Jeannette Song
4 - Selling Fashionable Products: Change Price or
Facilitate Learning?
Yufei Huang, PhD Student, University College London, Gower
Street, London, United Kingdom, yufei.huang.10@ucl.ac.uk,
Onesun Steve Yoo, Bilal Gokpinar, Chris Tang
We consider a manufacturer and a group of buyers who share a scarce but
expensive-to-build capacity over a finite period. Each member has private historydependent demand information and makes unverifiable investments. Because of
the high uncertainty, achieving supply chain efficiency while sustaining under a
dynamic environment is challenging for the partnership. We construct a
membership agreement that enforces efficient capacity allocation and investments
by introducing a novel breach remedy.
Firms selling new fashionable products are shifting their focus away from pricing
and towards facilitating the learning process for customers. To understand this
phenomenon, we present a stylized model with pricing and three channels
through which customers learn. We find that for new fashionable products,
facilitating learning can lead to greater profit than variable pricing. Moreover
when firms facilitate learning, variable pricing has only a marginal effect on firm
profits.
3 - The Perils of Sharing Information in a Trade-association
Noam Shamir, Assistant Professor, Tel-Aviv University, Haim
Levanon, Tel-Aviv, Israel, nshamir@post.tau.ac.il, Hyoduk Shin
■ TD52
Studying the incentives of a group of retailers, organized as a trade association, to
exchange forecast information, we compare between two industry policies:
exclusionary and non-exclusionary information sharing. Although nonexclusionary policy has been advocated to promote information sharing, we show
the opposite can happen and explain the reason.
52-Room 107A, CC
Social Media and Internet Marketing
Sponsored Session
4 - Aligning Incentives in Omni-channel Sale
Elnaz Jalilipour Alishah, PhD Candidate, University of
Washington, Seattle, Foster School of Business, Mackenzie Hall
358, Seattle, WA, 98195-3200, United States of America,
jalilipo@uw.edu, Yong-Pin Zhou, Jingqi Wang
Chair: Michael Trusov, University of Maryland,
3454 Van Munching Hall, College Park, MD, United States of America,
mtrusov@rhsmith.umd.edu
1 - Attribution Metrics and Return on Keyword Investment in Paid
Search Advertising
Hongshuang Li, Indiana University, Bloomington, IN,
United States of America, lhshruc@gmail.com, Siva Viswanathan,
Abhishek Pani, P.k. Kannan
We consider a retailer with both online and offline channels. While the online
store exerts costly effort to attract customers, the offline store handles inventory
for both locations – including fulfillment of online orders. We study how the
retailer should appropriately credit both channels to align their incentives.
In this paper, we analyze the impact of the attribution metric used for imputing
conversion credit to search keywords on the overall effectiveness of keyword
investments in search campaigns. We model the relationship among the
advertiser’s bidding decision for keywords, the search engine’s ranking decision
for these keywords, and the consumer’s click-through rate and conversion rate on
each keyword, and analyze the impact of the attribution metric used on the
overall return-on-investment of paid search advertising.
■ TD51
51-Room 106B, CC
Innovative and Entrepreneurial OM
Sponsored Session
2 - Controlling for Self Selection Bias in Customer Reviews
Leif Brandes, University of Warwick, Coventry, CV4 7AL,
United Kingdom, Leif.Brandes@wbs.ac.uk, David Godes,
Dina Mayzlin
Chair: Onesun Steve Yoo, University College London, Gower Street,
London, WC1E 6BT, United Kingdom, o.yoo@ucl.ac.uk
1 - Pricing and Capacity Planning for Flexible Consumption
Sanjiv Erat, UCSD, Gilman Drive, La Jolla, CA,
United States of America, serat@ucsd.edu, Sreekumar Bhaskaran,
Rajiv Mukherjee
Customers frequently use user online reviews as a valuable information resource
before making a purchase. This observation has motivated a large number of
empirical studies, and it is now a well-established finding that customer online
reviews impact product sales. However, one possible criticism of online user
reviews as a source of information is the self-selection inherent in the review
process. That is, consumers self-select into choosing whether to review a product,
which suggests that reviews may be prone to the extremity bias: the distribution
of reviews may be more polarized than the true preference distribution . This of
course implies that posted review valence may not always provide an unbiased
representation of customers’ true product experiences. We provide survey
evidence that demonstrates that customers who post an online review tend to
have more extreme opinions than customers who never post a review. We
hypothesize that the consumers who have more extreme opinions post their
reviews quicker, while the consumers with more moderate opinions may take
longer to post a review, which implies that in the limit some consumers with
moderate opinions may never post a review. One implication of this is that
reviews that arrive after a long time lapse are more similar to the opinions of the
non-responders. Hence, a firm that is able to observe the time lapse between the
experience and the review should be able to calculate the valence of reviews in a
way that corrects for the non-response bias. That is, we suggest how to correct for
the extremity bias by taking into account the latency of response data. To test our
hypotheses, we use a new dataset from a large online travel portal. Overall, we
have detailed information on 1.26 million bookings and 2.75 million reviews over
the complete history of the firm (twelve years). Because we observe hotel
bookings and review provision behavior at the individual customer level, we
know for each customer the exact duration between her last travel day and the
day that she provided the review. Based on our empirical results, we show how
customer self-selection across time impacts her review behavior and suggest a
method for controlling for this bias.
Motivated by the emergence of flexible consumption opportunities - such as the
rollover cellphone plans offered by many mobile providers - we study a firm’s
pricing and capacity planning decision when the timing of consumption is a
choice variable for consumers. Subsequently, we explore the effect of
heterogeneity in consumer preferences and its effect on a firm’s decision of how
much flexibility to offer.
2 - Improving Supply Chain Compliance using Buyer Consortiums
Prashant Chintapalli, Anderson School of Management,
University of California, Los Angeles, CA, 90095, United States of
America, prashant.chintapalli.1@anderson.ucla.edu,
Kumar Rajaram, Felipe Caro, Chris Tang
Motivated by the Accord on Fire and Building Safety in Bangladesh we study the
effectiveness of buyer consortiums. We show that a consortium can increase
factory compliance and improve the buyers’ profits, though possibly at the
expense of the supplier. We also study the conditions under which a buyer should
join the consortium and characterize the settings in which the whole supply chain
is better off.
3 - Startup as a Process: Increase Your Chances of Success via a
Just-in-time Approach
Christophe Pennetier, PhD Student, INSEAD,
1 Ayer Rajah Avenue, Singapore, 138676, Singapore,
Christophe.Pennetier@insead.edu, Karan Girotra,
Serguei Netessine
Using a unique and novel dataset, we study the success of startups modeled as a
process: what is the best configuration for batches of funding cash –in terms of
size and frequency– to exit successfully? Our results suggest that founders should
not be obsessed by the amount of money they raise in any single round. It is
better to raise small batches more often than the other way around.
363
TD53
3 - Deal or No Deal? The Quality Implications of Online Daily Deals
and Competition
Jorge Mejia, University of Maryland, Robert H. Smith School of
Business, College Park, MD, United States of America,
jmejia@rhsmith.umd.edu, Anand Gopal, Michael Trusov
4 - Is Non-linear Pricing Contract Always Better than Linear
Pricing Contract?
Guangwen Kong, University of Minnesota, 111 Church Street SE,
Minneapolis, MN, 55414, United States of America,
gkong@umn.edu, Tony Haitao Cui
Consumers use online reviews to inform purchasing decisions about many
products / services. Moreover, online daily deals have become an important part
of the marketing mix for merchants. The objective of this study is to understand
the effect of daily deals on consumers’ quality perceptions, expressed through
online reviews and investigate potential moderators for this effect, such as
merchant characteristics and competition. We combine online reviews for
restaurants from Yelp with data from online deals in a major American
metropolitan area. We find that online deals have a significant negative effect on
online reviews. Additionally, this effect is moderated by certain merchant
characteristics such as price point and restaurant age. We also find that the
reviews of merchants who do not offer deals are affected by nearby deal
competition. We replicate our empirical findings by conducting three lab studies
using subjects from MTurk and find consistent results, thus showing robustness.
We study supply chain contracts with consideration of information sharing and
bounded rationality. We examine a dyadic supply chain where a supplier with
more accurate demand information sells products to a bounded rational retailer.
The research suggests that the supplier can be better-off by using a linear pricing
contract than adopting a buy-back contract. The supplier either shares
information with the retailer or help improve the retailer’s bounded rationality
but not both in equilibrium.
■ TD54
54-Room 108A, CC
Meta-algorithms: From Algorithm Tuning and
Configuration to Algorithm Portfolios
4 - Swayed by the Numbers: The Consequences of Displaying
Review Counts in Purchase Decisions
Jared Watson, University of Maryland, College Park, MD, United
States of America, jwatson@rhsmith.umd.edu, Michael Trusov,
Anastasiya Pocheptsova
Cluster: Tutorials
Invited Session
Chair: Meinolf Sellmann, IBM, Yorktown Heights, NY,
United States of America
1 - Meta-algorithms: from Algorithm Tuning and Configuration to
Algorithm Portfolios
Meinolf Sellmann, IBM, Yorktown Heights, NY,
United States of America
Online retailers often display customers’ review to aid consumers’ decisionmaking. While prior literature postulates that an increase in review counts leads
to an increase in consumers’ purchase intentions, the authors find an important
corollary: holding purchase intentions constant, revealing a small review count
systematically biases consumers’ preferences between choice options. Further,
withholding review count information increases purchase intention relative to a
small review count. These findings are contrasted with current retailer practices of
revealing small review count information.
Efficiency and accuracy are of primary concern when developing analytics
solutions in OR. Typically, there is more than one possible algorithmic approach
and none dominates the others. Moreover, algorithms usually have implicit or
explicit parameters that greatly affect performance. Meta-algorithmics focuses on
the development of effective automatic tools that tune algorithm parameters and,
at runtime, choose the approach best suited for the given input. Here we
summarize the lessons learned when devising such tools.
■ TD53
53-Room 107B, CC
Inventory and Information Sharing
Sponsored Session
■ TD55
Chair: Enno Siemsen, Associate Professor, University of Minnesota,
321 19th Ave S, Minneapolis, MN, 55455, United States of America,
siems017@umn.edu
1 - Decision Dependent Bounded Rationality in Dual Sales
Channel Management
Ozalp Ozer, The University of Texas at Dallas, 800 West Campbell
Road, Richardson, TX, United States of America,
oozer@utdallas.edu, Kay-Yut Chen
55-Room 108B, CC
Data Envelopment Analysis (DEA)
Contributed Session
Chair: Samir Srairi, Ministry of Higher Education and Scientific
Research, 14 Avenue de Tunis, Arian, 2080, Tunisia,
srairisamir3@gmail.com
1 - Combination of Hybrid Two Level DEA with SVM for Indicator
Weighting in Financial Failure Prediction
Chao Huang, Southeast University, Xuan Wu District, Sipailou
No.2, Nanjing, 210096, China, huangchao@seu.edu.cn
We experimentally study behaviors in a dual sales channel in which a
manufacturer sells through his direct channel and an independent retailer. The
channels compete on demand. The manufacturer sets the wholesale price for the
retailer, and also delivery times for customers in his direct channel. The retailer
decides on its inventory level. We show and discuss why bounded rationality
differs, in the same subject pool, across three decisions and model the behavior as
quantal response equilibrium.
A new WPF two level DEA is proposed to identify the bankruptcy firms by
constructing a worst-practice frontier. Combining traditional BPF two level DEA
and WPF two level DEA, a hybrid model is put forward as a tool for cooperates
financial failure prediction. To improve the accuracy, a new indicator weighting
method based on SVM is also proposed. The empirical results show that the
proposed hybrid method has excellent bankruptcy prediction ability.
2 - Behavioral Inventory Sharing
Enno Siemsen, Associate Professor, University of Minnesota, 321
19th Ave S, Minneapolis, MN, 55455, United States of America,
siems017@umn.edu, Hui Zhao
2 - A Cross-Dynamic Evaluation of Warehouse Operations
Jose Humberto Ablanedo Rosas, Associate Professor, University of
Texas at El Paso, 500 W University Avenue, Marketing &
Management Department, El Paso, TX, 79968, United States of
America, jablanedorosas2@utep.edu, Faruk Arslan
The benefits of aggregating demand for reducing required safety stock
investments in supply chains are well known. Yet if decision makers are
decentralized and keep separate stockpiles of inventory, these benefits can only be
reaped if they agree to transship their inventory to others. Using behavioral
experiments, we explore the conditions under which decision makers share
inventory, and the implication of inventory sharing on initial order quantities.
We report a cross-dynamic comparison of distribution centers worldwide. This
problem has been analyzed with the Malmquist productivity index; we introduce
an approach based on cross-efficiency assessment which eliminates the drawbacks
of traditional cross-efficiency methods. A comparison between both approaches is
discussed and managerial recommendations for decision makers are derived.
3 - Communication Strategies in Assembly Systems:
An Experimental Investigation
Jud Kenney, McGill University, Bronfman Building, Montreal,
Canada, jud.kenney@mail.mcgill.ca, Jim Engle-Warnick,
Saibal Ray
3 - Performance Evaluation of Dynamic Supply Chain using Dea
Somayeh Mamizadeh-Chatghayeh, Asia Business Clusters &
Networks Development (ABCD) Foundation Cooperation,
Tehran, Tehran, Iran, somayeh_mamizadeh@yahoo.com,
Abbas Ali Noura
This study investigates how supply chain partners in an assembly system react to
three different strategies of communicating supply risk. Our behavioral
experiment uses a minimum game to model suppliers deciding on the amount of
capacity to build when facing certain end customer demand, but uncertain supply
from their peers. We find effective communication strategies can significantly
improve performance and such improvements are more significant under higher
critical ratios.
One of the main researches in supply chain management is to improve the overall
efficiency based on dynamic performance of supply chain. In this paper, by
developing the basic dynamic model, different methods for evaluating the supply
chain are studied. Dynamic Data Evolution Analysis as an efficient tool is new
research focus in supply chain benchmarking. We develop DDEA models that can
be evaluating the overall efficiency of supply chains and subsystems.
364
TD57
■ TD57
4 - On Integrating DEA and AHP for the Facility Layout Design in
Manufacturing Systems
Toloo Mehdi, Technical University of Ostrava, Ostracew,
Czech Republic, mehdi.toloo@vsb.cz
57-Room 109B, CC
Modeling the Economics of Low-Carbon
Power Systems
Analytic hierarchy process (AHP) is a decomposition multiple-attribute decision
making (MADM) method, which can represent human decision making process
and help to achieve better judgments based on hierarchy, pair-wise comparisons,
judgment scales, allocation of criteria weights and selection of the best alternative
from a finite number of variants by calculation of their utility functions. Data
envelopment analysis (DEA) is a well-known non-parametric method to evaluate
the performance of a set of homogeneous decision making units (DMUs) with
multiple inputs and multiple outputs.
Sponsored Session
Chair: Todd Levin, Energy Systems Engineer, Argonne National
Laboratory, 9700 S Cass Ave, Lemont, IL, 60439,
United States of America, tlevin@anl.gov
1 - Revenue Sufficiency and Resource Adequacy in Systems with
Variable Generation Resources
Todd Levin, Energy Systems Engineer, Argonne National
Laboratory, 9700 S Cass Ave, Lemont, IL, 60439,
United States of America, tlevin@anl.gov, Audun Botterud
5 - The Efficiency of Tunisian Universities: An Application of a
Two-Stage Dea Approach
Samir Srairi, Ministry of Higher Education and Scientific
Research, 14 Avenue de Tunis, Arian, 2080, Tunisia,
srairisamir3@gmail.com
This paper examines the efficiency of eleven universities in Tunisia during the
period 2009-2013. Regression analysis suggests that a higher share of professors, a
higher number of women in academic staff and a better quality of student in
secondary education improve the efficiency of the university.
An efficient MIP framework is applied to analyze the impact of increasing wind
power capacity on generator profitability. The model is executed with hourly time
steps on a test case that approximates the ERCOT system for a range of wind
capacity levels. We analyze three market policies that support resource adequacy
and find that some additional market incentives may be required to ensure long
term revenue sufficiency and resource adequacy in systems with significant
variable energy resources.
■ TD56
2 - An Approximate Model for Scheduling Energy and Reserve in
Renewable-Dominated Power Systems
Miguel Carrion, Universidad de Castilla - La Mancha,
Av. Carlos III, s/n, Toledo, Spain, miguel.carrion@uclm.es,
Rafael Zarate-miñano
56-Room 109A, CC
Project Selection, Evaluation and Collaboration
Invited Session
Since most of renewable energies are non-dispatchable, an appropriate schedule
of reserves in renewable-dominated power systems is crucial. We propose an
alternative formulation that co-optimizes energy and reserve considering the
uncertainty involved in demand and renewable production. This formulation
requires a significantly smaller number of variables and constraints than the
classical stochastic economic dispatch problem. The proposed formulation is tested
in a realistic case study.
Chair: Yaozhong Wu, National University of Singapore, NUS Business
School, Singapore, Singapore, yaozhong.wu@nus.edu.sg
1 - An Experiemental Study of Idea Selection Process
Zhijian Cui, Assistant Professor of Operations and Technology
Management, IE Business School, Calle de Maria de Molina 12,
Madrid, 28006, Spain, Zhijian.Cui@ie.edu, Shijith Payyadak,
Dilney Gonçalves
3 - Hydroelectric Bid Optimization under Uncertainty
Andy Philpott, University of Auckland, Engineering Science
Department, Private Bag 92019, Auckland, 1025, New Zealand,
a.philpott@auckland.ac.nz, Faisal Wahid, Frederic Bonnans,
Cedric Gouvernet
In this study, we design several online experiements to compares the efficacy of
two commonly observed processes of idea selection: scoring vs. ranking. In
scoring process, subjects are asked to evaluate the quality of each idea and give a
score while in ranking process, subjects are asked to only rank the ideas according
to their prefereces. We find that the choice of idea selection process depends on
some contextual factors.
We consider the problem faced by the operator of a cascade of hydroelectric
generating plants offering energy to a wholesale electricity pool market to
maximize revenue. Both energy prices and uncontrolled inflows to the reservoirs
of the cascade are assumed to be stochastic. We describe a stochastic dynamic
programming model that generates an optimal offer for the next period given
current observed prices. This is solved using SDDP when value functions are
concave or MIDAS when they are not.
2 - Overvaluation of Process Innovation Ideas
Fabian Sting, Erasmus University Rotterdam, Rotterdam,
Netherlands, fsting@rsm.nl, Christoph Fuchs, Maik Schlickel
Ideas by employees are a vital source for innovation. But are such ideas
overvalued by their creators? If so, which ideas in particular? Drawing on a
unique data set that comprises the generation, election, and implementation of
process improvement ideas of an automotive supplier, we identify antecedents of
overvalued ideas. Overvaluation is greater for ideas generated by higher-level
employees, collaboratively versus individually, and by employees with previously
lower ideation success.
4 - Optimal Timing to Invest, Mothball, Reactivate, and
Decommission a Coal Power Plant
Paul Rebeiz, Doctoral Candidate In Operations Mangement,
UCLA Anderson School of Management, 110 Westwood Plaza,
Los Angeles, CA, 90025, United States of America,
paul.rebeiz.1@anderson.ucla.edu, Christian Blanco
3 - Project Evaluation and Selection via Risk-adjusted Net
Present Value
Nicholas G. Hall, The Ohio State University, Fisher College of
Business, Columbus, OH, United States of America,
hall.33@osu.edu, Zhixin Liu, Wenhui Zhao
Transitioning to a low-carbon economy will require most coal power plants to be
replaced by other sources of generation such as wind and solar. We present a
dynamic program to solve for the optimal price signals to invest, mothball,
reactivate, and decommission a coal power plant. We find that our results are
consistent with current industry trends. We conclude with some insights on the
effect of renewable energy policy on mothballing and retiring a coal plant.
We consider a project with risk that declines over time as its tasks are completed,
as reflected in a declining discount rate. The objective is to maximize the NPV of
the project. This problem is highly nonlinear, since the discount rate at any point
in time is a function of previous scheduling decisions. We solve this model and
show that risk-adjusted NPV varies significantly from traditional NPV, and that the
use of the risk-adjusted measure significantly improves project selection decisions.
4 - Resource Competitions for Research Projects
Pascale Crama, Singapore Management University, 50 Stamford
Road, Singapore, 178899, Singapore, pcrama@smu.edu.sg, Anand
Nandkumar, Reddi Kotha
Academic research is funded by governments, but is often seeded through grants
from university administered research funds (UARF) and other charitable
institutions. We compare the effectiveness of UARF and other sources of funds in
obtaining subsequent federal funding and value creation. We build a
parsimonious model that can explain the superior productivity of UARF funding
and make recommendations on the ideal way to organize UARF funding.
365
TD58
■ TD58
Natural disasters such as Hurricane Sandy have seriously disrupted the power
grids. To increase the resilience of a distribution system under natural disaster
attacks, we propose a resilient distribution network design model considering
hardening and distributed generation unit placement to minimize the load
shedding under worst-case natural disaster attacks.
58-Room 110A, CC
New Insights on Electricity Markets with
Uncertain Supply
2 - Electric Resource Optimization with High Penetration Renewables
and Varying Reliability Measures
Cynthia Bothwell, Student, Johns Hopkins University, 117
Meridian Lane, Towson, MD, 21286, United States of America,
cdbothwell@gmail.com, Calvin Wood
Sponsored Session
Chair: Juan M. Morales, Associate Professor, Technical University of
Denmark, Matematiktorvet, Building 303b, 008, Kgs. Lyngby, 2800,
Denmark, jmmgo@dtu.dk
1 - The Benefits of Sharing Reserves Between Countries:
A Case Study of the European Electricity System
Kenneth Van Den Bergh, PhD Researcher, KU Leuven,
Celestijnenlaan 300 Box 2421, Leuven, Belgium,
kenneth.vandenbergh@kuleuven.be, Robin Broder Hytowitz,
Benjamin Hobbs, William D’Haeseleer, Erik Delarue
As intermittent wind and solar energy resources increase in use throughout the
electricity sector, techniques to assess system reliability are evolving. The
optimization of investment in new capacity resources changes as a result of the
reliability criteria applied to the system. This work overviews for policy and
decision makers the tradeoffs between reliability criteria and generation
investment with high penetrations of intermittent renewables for capacity
planning and market design.
3 - Scheduling Energy Storage Resources to Provide
Multiple Services
Johanna Mathieu, Assistant Professor, University of Michigan,
1301 Beal Ave, Ann Arbor, Mi, 48109, United States of America,
jlmath@umich.edu, Goran Andersson, Olivier Megel
Reserves are scheduled day-ahead in order to deal with forecast errors from wind
and sun. This study examines reserve coordination in time and space for the
Central European electricity system. Four scenarios involving various degrees of
coordination for reserves are simulated. A large-scale unit commitment model is
used to simulate the electricity markets. The study indicates savings in reserve
allocation costs of up to 90% with increasing degree of coordination.
Most energy storage devices in power systems are only partially used most of
time, and so they could also be used to help balance electricity supply and
demand. The challenge is how to allocate their energy and power capacities to
different services given uncertainty from multiple sources. We formulate the
scheduling problem and apply both stochastic dynamic programming and
stochastic dual dynamic programming to several case studies, and compare
performance and computational complexity.
2 - Effect of Ramping Pricing Scheme in Systems with High Wind
Energy Penetration
Yves Smeers, Professor Emeritus, Université Catholique de
Louvain, Voie du Roman Pays, 34, Louvain-la-neuve, B-1384,
Belgium, yves.smeers@uclouvain.be, Sebastian Martin
4 - Bidding Models for Price-responsive Loads in Electricity Markets
Javier Saez-gallego, PhD Candidate, Technical University of
Denmark, Matematiktorvet Building 303B, 019, Kgs. Lyngby,
2800, Denmark, jsga@dtu.dk, Juan M. Morales, Marco Zugno
We consider the continuous version of an unit commitment problem with wind
penetration, and subject to ramping constraints. The optimization problem
assumes that ramping providers are priced at opportunity cost. We explore the
impact of having different pricing schemes for ramping. We use a
complementarity formulation in several versions that differ to reflect different
policy proposals based on pricing schemes for ramping.
This paper presents a data-driven approach to estimate the parameters of the
market bid that best represents the stochastic and dynamic behavior of a pool of
price-responsive consumers. The proposed methodology is based on inverse
optimization and is able to leverage exogenous information, besides the electricity
price, to partly explain the parameters of the bid. We use data relative to the
Olympic Peninsula project to asses the performance of the proposed method.
3 - A Stochastic Electricity Market Clearing Formulation with
Consistent Pricing Properties
Victor M. Zavala, Computational Mathematician, Argonne
National Laboratory, 9700 South Cass Avenue, Argonne, IL,
60439, United States of America, vzavala@mcs.anl.gov
Deterministic clearing formulations introduce arbitrary distortions between dayahead and real-time prices that bias economic incentives. We analyze a stochastic
clearing formulation in which the social surplus function induces absolute value
penalties between day-ahead and real-time quantities. We prove that the
formulation yields price distortions that are bounded by the bid prices and we
prove that day-ahead quantities and flows converge to the medians of real-time
counterparts.
■ TD60
4 - On the Inefficiency of the Merit Order in Forward Electricity
Markets with Uncertain Supply
Juan M. Morales, Associate Professor, Technical University of
Denmark, Matematiktorvet, Building 303b, 008, Kgs. Lyngby,
2800, Denmark, jmmgo@dtu.dk, Salvador Pineda, Marco Zugno
Chair: Huan Liu, PhD, Xi’an Jiaotong University, 28# Xianning West
Road, Shaanxi Province, Xi’an, 710049, China, liuhuan-look@163.com
1 - Rapid Analysis of Attentional Processes While Looking at Print
Advertisements Based on Eye Tracking
Hirotaka Aoki, Dr., Tokyo Institute of Technology, 2-12-1-W9-75,
Oh-Okayama, Meguro-Ku, Tokyo, 152-8552, Japan,
aoki.h.ad@m.titech.ac.jp
60-Room 111A, CC
Performance Measurement
Contributed Session
We derive analytically the dispatch rule for a stylized power system with infinite
transmission capacity under a stochastic market-clearing mechanism. We provide
conditions for this clearing procedure to break the merit order and for virtual
bidding to ensure maximum market efficiency under a classical merit-order
dispatch. Finally, we provide a reinterpretation of these two market-clearing
procedures as members of a broader family that allows for marked-up forward
production costs.
This paper develops an eye tracking-based analysis framework for attentional
processes during viewing print advertisements. The framework consists of a
scheme for classification of information in advertising and principles for data
interpretation from attentional processes perspectives. Based on a case study in
which 20 consumers’ data during looking at insurance advertisements were
collected, the potentials of our framework as well as implications for effective
advertising design are discussed.
■ TD59
2 - Destructive Testing Gauge Capability Analysis
David Kim, Professor, Oregon State University, 204 Rogers,
Corvallis, OR, 97331, United States of America,
david.kim@orst.edu, Xinyu Luo
59-Room 110B, CC
Optimal Design and Operation of Smart
Electrical Grids
This research examines the current state-of-the-art in gauge capability analysis for
destructive testing. Results are then presented that extend the specific destructive
testing situations where gauge repeatability can be estimated.
Sponsored Session
3 - Project Timeliness or Project Effectiveness:
Which One is Sacrificed?
Asghar Afshar Jahanshahi, Postdoc, Pontifical Catholic University,
Av. Vicuña Mackenna 4860. Macul, Santiago, Chile,
asghar@ing.puc.cl, Khaled Nawaser
Chair: Baosen Zhang, University of Washington, 185 Stevens Way,
Seattle, WA, United States of America, zhangbao@uw.edu
1 - Design of Resilient Distribution Network Against Natural
Disasters: A Robust Optimization Approach
Bo Zeng, Assistant Professor, University of South Florida, Tampa,
4202 E. Fowler Avenue, Tampa, Fl, 33620, United States of
America, bzeng@usf.edu, Wei Yuan, Feng Qiu, Chen Chen,
Jianhui Wang
There is a consensus among scholars that real options reasoning is crucial for
project performance under conditions of high environmental uncertainty.
However, few empirical studies have confirmed this claim. Our longitudinal
analysis of 101 electronic commerce projects, drawn from new technology
ventures, indicated the differential effects of real options reasoning on project
performance under conditions of high environmental state, effect and response
uncertainty.
366
4 - A New Framework for Sustainability Measurement
Anthony Afful-Dadzie, University of Ghana, P.O. Box LG 78,
Legon, Accra, Ghana, atosarsah@gmail.com
TD63
4 - Maximizing Sustainability of Ecosystem Model through
Socio-economic Policies
Urmila Diwekar, President, Vishwamitra Research Institute, 2714
Crystal Way, Crystal Lake, IL, 60012, United States of America,
urmila@vri-custom.org, Kirti Yenkie, Rohan Doshi,
Pahola Benevades, Heriberto Cabezas
This presentation explores a new sustainability measurement and scoring system
for assessing the efforts of organizations at meeting sustainability targets. Based
on TOPSIS, the proposed measurement and scoring system incorporates all three
sustainability dimensions and enables the establishment of a threshold below
which an organization is considered to have failed a sustainability test. A timeindependent threshold is also introduced to help compare performance over time.
Current practices in natural resources consumption are unsustainable and may
eventually lead to ecosystem extinction. This paper uses a simple mathematical
model of an integrated ecological and economic system representing our planet’s
sectors. The aim of the project is to maximize the sustainability of this system,
using Fisher Information as a measure of sustainability, and derive socioeconomic policies using multivariable optimal control techniques.
5 - Government Intervention and Technovation Performance:
An Empirical Study of Soes from Mainland China
Huan Liu, PhD, Xi’an Jiaotong University, 28# Xianning West
Road, Shaanxi Province, Xi’an, 710049, China,
liuhuan-look@163.com, Jiannan Wu
■ TD62
We use a comprehensive provincial-level panel data set of 30 provinces during
2005-2012. Our results show that project funding and tax break at the provincial
level have no impact on new products sales. Project funding has a negative impact
on invention patents, by constrast, tax break has a positive impact on invention
patents. The interaction term of project funding and tax break has an invert Ushaped relationship with new products sales, but it has no impact on invention
patent.
62-Room 112A, CC
Optimization on Power Grid Application
Cluster: Energy Systems: Design, Operation, Reliability
and Maintenance
Invited Session
Chair: Chaoyue Zhao, Oklahoma State University, 322G Engineering
North, Stillwater, OK, United States of America,
chaoyue.zhao@okstate.edu
1 - Risk-based Admissibility Assessment of Wind Generation
Integrated into a Bulk Power System
Cheng Wang, Tsinghua University, 3-211, West Main Building,
Beijing, China, shlwangcheng2008@163.com, Feng Liu, Wei Wei,
Jianhui Wang, Shengwei Mei
■ TD61
61-Room 111B, CC
Environmentally Responsible
Operations Management
Sponsored Session
In this talk, a risk-based admissibility assessment approach is proposed to
quantitatively evaluate how much wind generation can be accommodated by the
bulk power system under a given UC strategy. Firstly, the operational risk brought
by wind generation is developed as an admissibility measure. Then a riskminimization model is established to mathematically characterize the admissible
region. Simulations demonstrate the effectiveness and efficiency of the proposed
methodology.
Chair: Arda Yenipazarli, Assistant Professor of Operations Management,
Georgia Southern University, COBA 2224, Statesboro, GA, 30460,
United States of America, ayenipazarli@georgiasouthern.edu
1 - Competitive Positioning and Pricing of Green Products with
Multiple Environmental Attributes
Arda Yenipazarli, Assistant Professor of Operations Management,
Georgia Southern University, COBA 2224, Statesboro, GA, 30460,
United States of America, ayenipazarli@georgiasouthern.edu
2 - Strong Formulations for Unit Commitment Problem
Kai Pan, PhD Student, University of Florida, 411 Weil Hall,
Gainesville, Fl, 32608, United States of America, kpan@ufl.edu,
Yongpei Guan
To address consumers’ sustainability-related product concerns, a thorough
approach to improving the environmental profile of one’s products is required.
Using one dimension of green may hide possible trade-offs and overlook the fact
that consumers’ preferences exhibit different orders in different green attributes.
We study a duopoly model that explicitly incorporates multiple environmental
attributes into the green product positioning and pricing, along with the trade-offs
among them.
In this talk, we will present the strong formulations for unit commitment problem
under different settings. Technical proofs are provided accordingly. Our
computational experiments verify the effectiveness of proposed strong
formulations.
3 - A Scalable Decomposition Method for the Two-Stage Stochastic
Unit Commitment Problem
Farzad Yousefian, Postdoctoral Research Associate, Penn State,
333 Logan Ave., Apt. 307, State College, PA, 16801, United States
of America, szy5@psu.edu, Wendian Wan, Uday Shanbhag
2 - Product Line Design: The Impact of Consumers’ Varied
Perceptions of Recycled Content
Monire Jalili, University of Oregon, 1208 University of Oregon,
Eugene, OR, United States of America, mjalili@uoregon.edu,
Nagesh Murthy, Tolga Aydinliyim
We consider a two-stage stochastic unit commitment problem modeled as a largescale mixed integer nonlinear optimization problem. The state-of-the art
commercial packages, e.g. CPLEX, do not scale with the number of the units and
scenarios. Motivated by the structure of the KKT system and employing the ideas
of Schur complements, we propose a multiphase primal-dual algorithm that scales
with the size of the scenarios. Preliminary simulation results are presents.
We consider a monopolist selling ordinary and green product versions to
consumers whose differential (dis)utility vary by consumer type, and is a function
of the firm’s quality decision (i.e., the amount of recycled content in the green
version.) We discuss how the optimal quality and pricing decisions drive demand
and profit and whether/when it is optimal for the firm to only offer the green
version (go completely green).
3 - Replenishment Decisions of Perishable Products under Price and
Emissions Sensitive Demand
Gokce Palak, Assistant Professor Of Operations Management,
Shenandoah University, Harry F. Byrd, Jr. School of Business,
Winchester, VA, 22601, United States of America, gpalak@su.edu
■ TD63
63-Room 112B, CC
Operations Management II
We extend economic lot sizing models for age dependent perishable products to
maximize profit and minimize emissions. This model captures the tradeoffs
between supplier and mode selection decisions, profits and emissions, and
transportation lead time and remaining shelf life of products. We analyze impacts
of price and emissions sensitive demand on the replenishment decisions.
Contributed Session
Chair: Xiaoyan Qian, PhD, The University of Auckland, 486 Parnell
Road, Auckland, New Zealand, x.qian@auckland.ac.nz
1 - Flexible Commitment Contract in the Presence of Goodwillsensitive Customers
Xiaoya Han, University of Science and Technology of China,
No. 96, Jinzhai Road, Hefei, China, xyhan@mail.ustc.edu.cn,
Yugang Yu
This paper focuses on a retailer’s dynamic decision problem: how to determine a
minimum commitment at the beginning of the planning horizon and periodically
variable order quantities to maximize its profit when facing goodwill-sensitive
customers. We obtain that the next-period goodwill decreases in the currentperiod one, and the goodwill monotonically converges to a constant steady-state
one over time. Moreover, we find that the steady-state goodwill may decrease in
minimum commitment.
367
TD64
2 - Flexible Capacity Management with Advanced Information
Julian Kurz, Chair of Logistics and Quantitative Methods in
Business Administration, University of Wuerzburg,
Stephanstrasse 1, Wuerzburg, 97070, Germany,
julian.kurz1@uni-wuerzburg.de
3 - Modeling Reference Dependence using
One-switch Independence
David Vairo, Virginia Commonwealth University, 2415 Krossridge
Road, N. Chesterfield, VA, 23236, United States of America,
vairodl@vcu.edu, Jason Merrick
We consider a maintenance service provider that overhauls technical equipment
for customers in a central facility. A flexible capacity control policy is developed
such that capacity costs and queue length-related holding costs are minimized.
We investigate three operating modes, each taking into account a different
amount of information (reactive/single-/multi-stage proactive modes). In the
proactive modes, advanced information regarding future jobs is utilized.
We present an application of multi-attribute one-switch independence to single
attribute gambles by modeling chance as an attribute, which models reference
dependence and shows it is equivalent to one-switch independence. The resulting
form obeys stochastic dominance while incorporating probabilistic sensitivity,
utility curvature, reference dependence, and loss aversion. The approach connects
single-attribute behavioral and multi-attribute prescriptive decision analysis.
3 - How to Take Advantage of Crowdsourcing to Collect New Ideas
about Product Innovation?
Wanjiang Deng, Huazhong University of Science and Technology,
School of Management, Luoyu Road 1037, Hongshan District,
Wuhan, 430074, China, dengwj01@foxmail.com, Shihua Ma
4 - Multiobjective Network Resilience Model with Parallel
Component Recovery
Nazanin Morshedlou, PhD Student, University of Oklahoma, 202
W. Boyd St., Room 424, Norman, OK, 73071, United States of
America, nazanin.morshedlou@ou.edu, Kash Barker
Crowdsourcing is gaining more and more attention in both practice and research
field. We stand on the position of the company who is going to propose a task
about product innovation on one online crowdsourcing platform, and investigate
its best strategies of both choice of platform and reward setting. We derive
solutions of the base model and then extend it to some detail aspects. Finally, we
discuss the managerial insights of our research.
This work introduces a multiobjective formulation that trades off investments to
enhance network resilience in the form of (i) strengthening link capacity
following a disruptive event to decrease vulnerability, and (ii) introducing
“parallel component at a time” recovery scheduling to improve recoverability.
Given the uncertainty associated with critical infrastructures, robust interval
optimization is used to solve the multiobjective formulation
4 - A Heuristic for Hospital Operating Theatre Scheduling
under Uncertainty
Milad Zafar Nezhad, Wayne State University, Industrial and
Systems Engineering Dep, Detroit, MI, 48202, United States of
America, fq3963@wayne.edu, Hossein Badri, Kai Yang
■ TD65
65-Room 113B, CC
Resource planning is one of the most important issues in healthcare operating
management. In this research a heuristic solution algorithm based on the shifting
bottleneck method is developed for hospital operating theatre scheduling when
some parameters are not deterministic. The developed algorithm is applied on
several instances to evaluate its applicability and performance.
Near Miss and Threshold Events and Their Influence
on Risk Perception and Behavior
5 - Contractual Coordination of Agricultural Cooperatives with
Quality Specifications
Xiaoyan Qian, PhD, The University of Auckland, 486 Parnell
Road, Auckland, New Zealand, x.qian@auckland.ac.nz
Chair: Florian Federspiel, IE Business School, Maria de Molina 12,
Bajo, Madrid, 28006, Spain, ffederspiel.phd2014@student.ie.edu
1 - Conceptualizing Perceptions of Near Misses
Richard John, Associate Professor, University of Southern
California, 3620 McClintock Ave., Dept. of Psychology, MC-1061,
Los Angeles, CA, 90266-1061, United States of America,
richardj@usc.edu, Jinshu Cui, Heather Rosoff
Sponsored Session
This talk examines how agricultural cooperatives can motivate farmers’ effort
when the market price depends on the quantity of high quality produce. We
assume that a quality premium is offered to farmers and that their pay-outs are
made progressively. We propose a two-stage stochastic model. The main findings
are conditions for when the supply chain can be coordinated, that effort is
motivated by the quality requirement, and that the progressive payment is
needed for coordination.
I will present a dynamic probabilistic model for describing and defining near miss
events. The model is useful for highlighting characteristics of a near miss that
determine the extent to which the event is perceived as a near miss.
Measurement of individual differences in perceptions of near misses will also be
discussed.
2 - Small Near-Misses: Too Weak of a Warning Signal
Robin Dillon-Merrill, McDonough School of Business,
Georgetown University, Washington, DC, 20057,
United States of America, rld9@georgetown.edu
■ TD64
64-Room 113A, CC
Optimization and Utility Theory
This research demonstrates that individuals too often evaluate near-misses as
successful events even when it was fortunate chance that prevented the same
event from resulting in a failed outcome. These weak signals of problems are then
overlooked as the warnings they could potentially be. This problem is exacerbated
as small near-misses accumulate over time, and decision makers increasing accept
more risk.
Sponsored Session
Chair: Hiba Baroud, Vanderbilt University, 400 24th Avenue South,
Nashville, TN, 37205, United States of America,
hiba.baroud@vanderbilt.edu
1 - A Multi-criteria Decision Analysis Approach for Importance
Ranking of Network Components
Yasser Almoghathawi, PhD Candidate, University of Oklahoma,
202 W Boyd St., Norman, OK, 73019, United States of America,
moghathawi@ou.edu, Kash Barker
3 - Learning from Threshold Events
Wenjie Tang, IE Business School, Calle de Maria de Molina 12,
Piso 5, Madrid, MA, 28006, Spain, Wenjie.Tang@ie.edu,
Steffen Keck, Matthias Seifert
Threshold events are events that are triggered when an observable underlying
random variable passes a known threshold. We suggest that when individuals
learn from past threshold events, their judgments depend strongly on whether
the realization of the underlying variable triggers the threshold event; the extent
to which the realization has been close to the threshold will be discounted.
Results from a laboratory experiment support our main hypothesis.
Analyzing network vulnerability is a key element of network planning and
preparing for a disruptive event that might impact the performance of the
network. Many importance measures have been proposed to identify and rank
the important components in a network to focus on preparedness efforts. We
integrate a number of flow-based importance measures with a multi-criteria
decision analysis technique, TOPSIS, highlighting how different weighting
schemes can lead to different rankings.
4 - The Experience of Near Miss Events under Ambiguity
Florian Federspiel, IE Business School, Maria de Molina 12,
Bajo, Madrid, 28006, Spain, ffederspiel.phd2014@student.ie.edu,
Matthias Seifert
2 - Modeling Uncertainty in Risk-preference Elicitation
Dharmashankar Subramanian, Research Staff Member, IBM
Research, 1101 Kitchawan Rd, Rte 134, Yorktown Heights, NY,
10598, United States of America, dharmash@us.ibm.com,
Debarun Bhattacharjya, Mengyang Gu
Near miss events are often clouded in ambiguity, allowing for hubris and
misattribution of what caused success or prevented failure. We investigate the
experience of near miss events, probabilistic events nearly resulting in loss
outcomes that do not materialize due to chance circumstances, and related
changes in risk perception and behavior under both ambiguity and unawareness.
We find that the near miss effect (an increase in risk taking behavior) principally
occurs under ambiguity.
Utility functions are used to model a decision-maker’s risk-preferences. However,
interactive elicitation of a precise utility function is fraught with many practical
challenges such as noise, inconsistency and bias in the responses to questions. In
this work, we provide a flexible model along with Bayesian analysis to calibrate
random utility functions and to quantify different sources of uncertainty. We
present both theoretical and numerical results.
368
■ TD66
TD68
2 - Designing a Biorefinery Supply Chain: a Real Case in
Navarre (Spain)
Adrian Serrano, Public University of Navarra, Pamplona
Spain, adrian.serrano@unavarra.es, Javier Belloso, Javier Faulin,
Alejandro G. del Valle
66-Room 113C, CC
Air Cargo
Sponsored Session
New alternative energy sources are spreading around the world to reduce
greenhouse gas emissions and oil dependence. Our paper proposes a procedure to
manage a biorefinery supply chain in Navarre (Spain) which involves, among
others, which farms are going to be harvested, when they are going to be
collected, and the storage levels. Moreover, a Facility Location Problem is solved
inside a MILP model. Promising results are obtained at both levels: strategic
(location) and operational (SCM).
Chair: Jose Quesada, Université Catholique de Louvain, Chausée de
Binche, 151, Mons, 7000, Belgium, jose.quesada@uclouvain-mons.be
1 - An Economic Analysis of the Air Cargo Problems in an Integrated
Supply Chain
Kwon Gi Mun, PhD Candidate, Rutgers University, SCM,
Rutgers Business School, 1 Washington Park, Newark, NJ, 07102,
United States of America, kwongimun@gmail.com,
Yoondong Jung, Yao Zhao, Endre Boros, Arim Park
3 - Train Dispatching Problem under Exact Travel Time Estimation for
a Double Track Rail System
Lance Fu, University of Southern California, Los Angeles, CA,
United States of America, luncefu@usc.edu, Maged Dessouky
In this model, we demonstrate an integrated forecasting approach to coordinate
ground and air transportation for a Korean air cargo company. Therefore, we
present expected benefits of this integrated approach compared to current
practice.
We consider the problem of dispatching trains through double track railway
system, where track segments have different speed limits. We take the train’s
dynamics into consideration, which differentiates our model from the previous
literature. The objective is to minimize the traveling time under no-deadlock and
no-collision constraints. We give a mixed integer programming (MIP) formulation
for the train dispatching problem. Also we provide certain conditions which can
ensure that there exists an optimal integer solution to relaxation of the MIP. A
local search based heuristic is also proposed to solve the problem. Simulation on
the railway system in Los Angeles County is conducted to verify the efficiency of
the proposed algorithms.
2 - A Multi-Stage Air Service Network Design Problem for an
Express Carrier
Yusuf Secerdin, University of Miami, 1251 Memorial Drive,
Department of Industrial Engineering, Coral Gables, FL, 33146,
United States of America, yusufsecerdin@miami.edu,
Murat Erkoc
4 - A Stochastic Programming Approach for Truckload Relay Network
Design under Demand Uncertainty
Zahra Mokhtari, Oregon State University, Corvallis, OR,
United States of America, mokhtarz@onid.oregonstate.edu,
Hector A. Vergara
We study the air service network configuration problem for a global express
carrier. We propose a multi-stage modeling framework for the company’s Central
and South America region by incorporating multiple service types in terms of
time commitments for the air network. The proposed approach consists of three
phases in which we formulate a hub location problem, generate feasible pick-up
and delivery routes and formulate the service network design problem using the
composite variable formulation.
This study addresses the problem of strategic relay network design for truckload
transportation under demand uncertainty and proposes a stochastic programming
model and solution algorithm. The solution methodology uses Sample Average
Approximation (SAA) to address a very large number of scenarios of demand
realization. The examined number of scenarios determines the trade-off between
optimality of the solutions obtained for the stochastic programming model and its
computational complexity. Numerical results on a set of instances of this problem
are presented along with areas for future research.
3 - An Adaptive Search Network for the Pickup and Delivery Problem
with Time Windows
Ferdinand Kiermaier, TU Munich, Arcisstr. 21, Munich, Germany,
ferdi.kiermaier@googlemail.com, Jonathan Bard, Markus M. Frey
We present an innovative “out-of-the-box” algorithmic framework coupling
existing heuristics with a learning-based network structure applicable to many
variants of the Pick-Up and Delivery Problem with Time-Windows (PDPWTW)
and, thus, for the Vehicle Routing Problem with Time-Windows. We show an
application to a real-world airport baggage and cargo transportation problem and
proove the effectiveness of our new approach by a comparison with state-of-theart solution algorithms for the PDPWTW.
■ TD68
68-Room 201B, CC
Resilience in Electricity Infrastructure Systems
4 - Express Air Network Design with Multi-Hub Flexible Connections
Jose Quesada, Université Catholique de Louvain, Chausée de
Binche, 151, Mons, 7000, Belgium, jose.quesada@uclouvainmons.be, Jean-sébastien Tancrez, Jean-charles Lange
Sponsored Session
Chair: Yong Fu, Associate Professor, Mississippi State University,
Starkville, MS, United States of America, fu@ece.msstate.edu
1 - Microgrids for Enhancing the Power System Resilience,
Reliability, & Economics
Mohammad Shahidehpour, Professor, IIT, 10 West 35th Street,
Suite 1600, Chicago, IL, 60616, United States of America,
ms@iit.edu
We present a model for the Air Network Design for the next day delivery within
an Express company. Most of the existing models rely on a pre-definition of
connections for each commodity through a specific hub. We present a model in
which we integrate the decision of connectivity simultaneously with the network
design. When two hubs are so close from each other that they can serve (almost)
the same nodes, the results show that savings can be obtained by taking both
decisions at the same time.
Microgrids form the building blocks of perfect power systems which promote the
use of real-time pricing and demand response for optimizing the distributed
control of electric power systems. This presentation will highlight some of the key
issues in the design and the operation of microgrids and discuss the role of recent
innovations and, in particular, the significance of smart grid applications to power
system operations and control.
■ TD67
67-Room 201A, CC
Topics in Transport I
2 - Mitigating Cascading Outages under Severe Weather using
Simulation-based Optimization
Jianhui Wang, Argonne National Laboratory, 9700 South Cass
Avenue, Building 221, Argonne, IL, 60439, United States of
America, jianhui.wang@anl.gov, Feng Qiu, Jie Xu
Sponsored Session
Chair: Zahra Mokhtari, Oregon State University, Corvallis, OR,
United States of America, mokhtarz@onid.oregonstate.edu
1 - A Hybrid Heuristic Method for the Compressed Natural Gas Truck
Routing Problem with Fueling Stations
Yihuan (Ethan) Shao, University of Southern California, Los
Angeles, CA, United States of America, yihuansh@usc.edu,
Maged Dessouky
In this work, we investigate cascading outage mitigation under severe weather
conditions. Since the cost function, expected cascading outage costs, cannot be
expressed as an explicit function of protection actions and system status, we
develop a power system security simulator to estimate the cascading outage costs
of given mitigation actions and use a simulation-based optimization approach.
We introduce the Compressed Natural Gas Truck Routing Problem with Fueling
Stations to model decisions to be made with regards to the vehicle routes
including the choice of fueling stations. A hybrid heuristic method is proposed,
which combines an Adaptive Large Neighborhood Search (ALNS) with a mixed
integer program. By solving a set of benchmark instances, we show the
effectiveness of the method. We also conduct experiments based on the data from
the Ports of Los Angeles and Long Beach.
369
TD69
3 - A Decentralized Decision Making System to Enable Resilient
Microgrid Clusters
Yong Fu, Associate Professor, Mississippi State University,
Starkville, MS, United States of America, fu@ece.msstate.edu
■ TD70
Microgrid has been proposed to ensure resilience in power systems. The microgrid
can treat connected neighboring microgrids as local energy buffers thus freely
forming a cluster to share, exchange, and aggregate site-generated energy. This
research proposes a decentralized decision making system to improve the
microgrid clusters’ resilience capability to power disturbances and extreme events,
consequently minimizing down-time for both consumers and the grid.
Sponsored Session
70-Room 202A, CC
Tutorial: Railroad Predictive Analytics
Chair: Aihong Wen, CSX, 500 Water St, Jacksonville, FL, 32202,
United States of America, aihong_wen@csx.com
1 - Railroad Data Mining Tutorial
Aihong Wen, CSX, 500 Water St, Jacksonville, FL, 32202,
United States of America, aihong_wen@csx.com, Jerry Kam
■ TD69
We will share the business use cases and modeling experiences in applying data
mining and big data techniques to railroad.
69-Room 201C, CC
Connected and Autonomous Vehicles I
Sponsor: TSL/Intelligent Transportation Systems (ITS)
Sponsored Session
■ TD71
71-Room 202B, CC
Chair: Yong Hoon Kim, Purdue University, United States of America,
kim523@purdue.edu
1 - Multi-agent Based Formation Control of Connected
Autonomous Vehicles
Yongfu Li, Chongqing University of Posts and
Telecommunications, Chongqing, China, laf1212@163.com,
Kezhi Li, Li Zhang, Srinivas Peeta, Xiaozheng He, Hong Zheng,
Taixiong Zheng
Transportation Planning II
Contributed Session
Chair: Hadi Farhangi, Research Assistant, Missouri University of
Science and Technology, 1870 Miner Cir, Rolla, MO, 65401,
United States of America, hfrhc@mst.edu
1 - NHTSA Cafe Compliance Cost Optimization
Yohan Shim, Sr. Analyst, AVL Scenaria Inc, 47603 Halyard Drive,
Plymouth, MI, 48170, United States of America,
yohan.shim@scenaria.com, Travis Tamez, Christopher Mollo,
Frederic Jacquelin
This study seeks to improve network throughput and reduce energy consumption
under V2V communications environment through formation control. A multiagent systems based formation control is proposed using consensus theory. We
analyze the formation of autonomous vehicles in longitudinal and lateral gaps
simultaneously. Numerical experiments illustrate the effectiveness of the
proposed method in terms of position and velocity consensus.
The United States National Highway Traffic Safety Administration (NHTSA) has
issued in August 2012 final rules and regulations for Corporate Average Fuel
Economy (CAFE) for model years 2017 and beyond. NHTSA sets national CAFE
standards under the Energy Policy and Conservation Act to improve fuel
economy for passenger cars and light trucks. We present a mathematical program
model and efficient heuristics to support vehicle manufacturer’s long-term
strategic decisions on CAFE credit utilization.
2 - Macroscopic Modeling of the Spatial-temporal Information Flow
Propagation Waves under Vehicle-to-Vehicle Communications
Yong Hoon Kim, Purdue University, West Lafayette, IN,
United States of America, kim523@purdue.edu, Srinivas Peeta,
Xiaozheng He
2 - Highway Cost Allocation for Vehicle Classes with Variable Traffic
Capacity Requirements
Saurav Kumar Dubey, PhD Student, Department Of Industrial
And Systems Engineering, University of Tennessee at Knoxville,
1615 Laurel Avenue, Knoxville, TN, 37916, United States of
America, skumardu@vols.utk.edu, Alberto Garcia-Diaz
This study proposes an integrated model consisting of integro-differential
equations to describe the information flow propagation process and a partial
differential equation to describe the traffic flow dynamics. It incorporates the
success rate of communication with distance and interference as a probability
density function, and provide a closed-form solution for the speed of the
information propagation wave. Numerical experiments are conducted to analyze
the performance of the proposed model.
A Highway Cost Allocation model with variable traffic capacity levels is developed
to distribute costs among vehicle classes. The discrete Aumann-Shapley value is
used to generate costs for all coalitions in the least core model. Rules for tie
breaking to get a unique allocation known as the nucleolus are discussed.
3 - Vehicle Trajectory Reconstruction under the Mixed Connected
Vehicle Environment
Feng Zhu, Purdue University, WEst Lafayette, IN,
United States of America, zhu214@purdue.edu,
Satish V. Ukkusuri
3 - Competition and Regulation of the Taxi Market with
Ride-sourcing Platforms
Liteng Zha, University of Florida, 365 Weil Hall, Gainesville, FL,
32611, United States of America, seuzha@gmail.com, Yafeng Yin
This paper sets out to reconstruct the trajectory of non-connected vehicles based
on the trajectory of connected vehicles only. The trajectory reconstruction
problem is formulated in the linear state-space modeling (SSM) framework,
where the state dynamics is captured by the simplified car following model. Next
a modified EM (Expectation-Maximization) algorithm is developed to obtain the
optimal estimation of the unknown trajectory and model parameters
simultaneously. The performance of the EM algorithm is tested and validated
through the simulation data.
Ride-sourcing platforms such as Uber and Lyft are eroding the traditional taxi
market. Despite of their attractiveness, controversies arise over the legality and
reliability of their services as well as the fairness of the competition to regular
taxis. This study offers a quantitative investigation of the taxi market with ridesourcing platforms and investigates its regulation strategies.
4 - Biobjective Efficient Driving of Electric Vehicles on an
Edge of a Network
Hadi Farhangi, Research Assistant, Missouri University of Science
and Technology, 1870 Miner Cir, Rolla, MO, 65401, United States
of America, hfrhc@mst.edu, Dincer Konur, Warren Vaz,
Umit Koylu
4 - A Mathematical Model to Locate Optimal Lane Changing Zone at
a Highway Off-ramp
Siyuan Gong, Illinois Institute of Technology, Chicago, IL,
United States of America, sgong1@hawk.iit.edu, Lili Du
This research seeks to locate an optimal temporal-spatial lane change zone around
a highway off-ramp. This zone will grant enough opportunities for vehicles to
proposed and validated by simulation experiments.so that the resulted time delay
can be minimized. A mathematical model combining traffic flow analysis is
conduct smooth lane change maneuvers before an off ramp
This study analyzes driving profile for an electric vehicle between two nodes, i.e.,
on an edge of a network. We consider acceleration, maximum speed, and
deceleration as decision variables. The energy consumed and the travel time
required are functions of these variables as well as the load carried and the
distance traveled. We investigate the energy and time minimizing driving profiles
through a bi-objective optimization approach.
370
■ TD72
TD74
2 - Generating and Comparing Pareto Fronts of Experiment Designs
to Account for Multiple Objectives
Byran Smucker, Assistant Professor, Miami University,
100 Bishop Circle, 311 Upham Hall, Oxford, OH, 45056,
United States of America, smuckerb@miamiOH.edu,
Yongtao Cao, Tim Robinson
72-Room 203A, CC
Design and Analysis of Data with Complex Structure
Sponsored Session
In many design scenarios the experimenter entertains multiple, conflicting
objectives. The Pareto approach to experiment design is to construct a set of
designs while explicitly considering trade-offs between criteria. The true Pareto
front is not known, which creates problems in assessing front quality, and existing
algorithms are inefficient, ineffective, or both. Here, we introduce an improved
measure of front assessment, and present a new algorithm to generate Pareto
fronts of designs.
Chair: Xinwei Deng, Assistant Professor, Department of Statistics,
Virginia Tech, 211 Hutcheson Hall, Blacksburg, VA, United States of
America, xdeng@vt.edu
Co-Chair: Ran Jin, Virginia Tech., Grado Department of Industrial and,
Systems Engineering, Blacksburg, VA, 24061, United States of America,
jran5@vt.edu
1 - Markov Switching Autoregressive Models with Applications in
Cell Biology
Ying Hung, Rutgers University, Piscataway, NJ,
United States of America, yhung@stat.rutgers.edu
3 - Cost Constrained ALT with Exponentially Changing
Stress Durations
David Han, University of Texas, One UTSA Circle, San Antonio,
TX, United States of America, David.Han@utsa.edu
When designing ALT, several variables such as the allocation proportions and
stress durations must be determined carefully because of constrained resources.
This talk discusses the optimal decision variables based on the popular optimality
criteria under the constraint that the total cost does not exceed a pre-specified
budget. A general scale family of distributions is considered to accommodate
different lifetime models for flexible modeling with exponentially decreasing
stress durations.
We will introduce a new framework based on Markov switching autoregressive
models for the analysis of experiments in cell biology.
2 - Sparse Particle Filtering
Yun Chen, University of South Florida, 4202 E. Fowler Ave.
ENB118, Tampa, FL, United States of America,
yunchen@mail.usf.edu, Hui Yang
4 - Integration of Computer and Physical Experiments for Improving
Predictive Inference
Arda Vanli, Associate Professor, Florida State University,
Tallahassee, FL, avanli@fsu.edu, Spandan Mishra
Wireless sensor network has emerged as a key technology for monitoring spacetime dynamics of complex systems. Distributed sensing gives rise to
spatially-temporally big data. Realizing the full potentials of distributed sensing
calls upon the development of space-time modeling of measured signals in
dynamically-evolving physical environment. This paper will present a new
approach of sparse particle filtering to model spatiotemporal dynamics of big data
in distributed sensor network.
A Bayesian predictive approach is developed to combine data from designed
experiments on physical process and computer predictions. Predictive distribution
of a regression model is used for inference on the outcome variable and issues
including predictive capability, model adequacy and sensitivity to prior
specifications are discussed. Applications from structural loss prediction and
quality control are presented for illustrations.
3 - Graphical Modeling with Functional Variables
Ran Jin, Virginia Tech., Grado Department of Industrial and,
Systems Engineering, Blacksburg, VA, 24061, United States of
America, jran5@vt.edu, Hongyue Sun, Shuai Huang
■ TD74
Graphical models are widely used to model variable relationship. Traditional
graphical models are mainly used to model scalar variables. In this paper, a
graphical model with functional variables is proposed. Functional regression
models, combined with sparsity-inducing norms, are applied for the graphical
modeling. A case study and simulation will be used to evaluate the proposed
method.
74-Room 204A, CC
Bayesian Applications in Industrial Statistics
Sponsored Session
4 - Optimal Design of Experiments for Generalized Linear Models
Abhyuday Mandal, University of Georgia, Department of
Statistics, 101 Cedar Street, Athens, GA, 30602-7952, United
States of America, amandal@stat.uga.edu, Liping Tong, Jie Yang,
Dibyen Majumdar
Chair: Refik Soyer, The George Washington University, 2201 G St NW,
Washington, DC, United States of America, soyer@gwu.edu
1 - What do Coin Tosses, Vessel Traffic Risk Assessment and Return
Time Uncertainty Have in Common?
Johan Rene Van Dorp, Professor, The George Washington
University, 800 22nd Street NW, Suite 2800, Washington, DC,
United States of America, dorpjr@gwu.edu, Jason Merrick
Generalized linear models have been used widely for modeling the mean
response both for discrete and continuous random variables with an emphasis on
categorical response. Here we find efficient designs in the context of several
optimality criteria, namely D-optimality, EW-optimality and Bayesian optimality.
Regular fractional factorials with uniform replications are often used in practice.
We show that these popular designs are often not optimal for binomial, Poission
and multinomial cases.
Via a coin toss argument we will advocate decision making under uncertainty in
vessel traffic risk assessment to be informed by relative risk comparisons by
highlighting the analogy of an accident potentially occurring in a traffic situation
with the toss of a biased coin. That same analogy is next used to demonstrate the
large uncertainty bands that result for average return times of accidents in this
context.
■ TD73
2 - Integrating Expert Judgement and Bayesian Analysis
Thomas A. Mazzuchi, Professor And Chairman,
George Washington University, Washington DC, DC, 20052,
United States of America, mazzu@gwu.edu
73-Room 203B, CC
Recent Advances in Analyzing Experiments
Sponsored Session
There is a growing need for marrying the fields of expert judgement and Bayesian
Analysis that is, using the Expert Judgement approach to define prior
distributions and for understanding the effects of the elicitation, codification and
combination on the prior distribution and subsequent posterior analysis. This
paper presents an investigation of the above for a simple model using the Classical
Model for Expert Judgment by Cooke (2001).
Chair: Rong Pan, Associate Professor, Arizona State University, P.O. Box
878809, Tempe, AZ, United States of America, rong.pan@asu.edu
1 - Performance of Standard Mixture Designs in Modeling
Ordinal Responses
Mickey Mancenido, Arizona State University, 699 S. Mill Ave.,
Tempe, AZ, United States of America, mmanceni@asu.edu,
Douglas Montgomery, Rong Pan
3 - An Augmented Simulation Approach for Bayesian Design of
Life Tests
Refik Soyer, The George Washington University, 2201 G St. NW,
Washington, DC, United States of America, soyer@gwu.edu
Mixture designs for the proportional odds model — the widely used model for
ordinal data — are lacking in literature. A viable surrogate are the standard
mixture designs for linear models with normal errors. We are interested in the
performance of the simplex-lattice with axial runs, simplex-centroid, computergenerated I-optimal, and the uniform space-filling designs when used in a
mixture study with an ordinal response.
In this talk we consider a Bayesian decision theoretic setup for optimal design of
life tests. More specifically, we consider use of augmented probability simulation
with a conjugate class of utility functions for design of life tests. We illustrate the
implementation of the approach in one and two stage designs.
371
TD75
4 - Bayesian Updating of Dirichlet Process Prior via Kernel Estimate
Ehsan Soofi, Univesity of Wisconsin, Lubar School of Business,
Milwaukee, WI, United States of America, esoofi@uwm.edu,
Neshat Beheshti, Jeffrey Racine
■ TD76
The standard nonparametric Bayesian approach uses multinomial proportions to
update the Dirichlet Process Prior (DPP). We use kernel-smoothed CDF instead of
the multinomial proportions for updating DPP. Applications include Bayesian
measures and inferences for distributional fit and for dependence of random
variables via the information measure of copula. The posterior mean of the
quantized entropy provides a Bayes estimate of the dependence.
Sponsor: Simulation
Sponsored Session
76-Room 204C, CC
Rare Event Simulation and Network Applications
Chair: John Shortle, George Mason University, 4400 University Dr.,
MSN 4A6, Fairfax, VA, United States of America, jshortle@gmu.edu
1 - Rare-event Simulation for Queues with Time-varying Arrivals
Ni Ma, Columbia University, 500 West 120th Street, Room 345,
nm2692@columbia.edu, Ward Whitt
■ TD75
75-Room 204B, CC
We show that the exponential tilting approach for rare-event simulation in the
GI/G/1 queue can also be applied to efficiently estimate the time-varying
periodic-steady-state probability of large delays in a Mt/GI/1 single-server queue
with periodic arrival rate function.
IBM Research Best Student Paper Award IV
Sponsored Session
2 - Green Simulation Designs for Repeated Experiments
Mingbin Feng, PhD Candidate, Northwestern University, 2145
Sheridan Rd, Rm C210, Evanston, IL, 60208, United States of
America, benfeng@u.northwestern.edu, Jeremy Staum
huangmh@ntu.edu.tw
huangmh@ntu.edu.tw
In many applications of simulation, such as in financial risk management,
experiments are usually repeated with similar inputs. In these cases simulation
outputs should be viewed as useful resource that should be recycled and reused to
improve the efficiency of subsequent experiments. We consider a periodic credit
risk evaluation problem in the KMV model and the numerical results show
improving accuracy over time, measured by mean squared error, as more and
more outputs are recycled.
2 - Can Objective Early Warning Scores and Subjective Risk
Assessments Predict Patient’s Hospital Length of Stay
and Mortality?
Nasibeh Azadeh-Fard,Phd Candidate, Virginia Tech,
544 Whittemore Hall, Virginia Tech, Blacksburg VA 24061,
United States of America, nasibeh@vt.edu, Jaime Camelio, Navid
Ghaffarzadegan
3 - A General Golf Course Simulation Tool: Keeping Delays Down and
throughput Up
Moonsoo Choi, Columbia University, Department of Industrial
Engineering, New York, NY, 10027, United States of America,
mc3983@Columbia.edu, Qi Fu, Ward Whitt
We describe a simulation tool for designing and managing golf courses. Group
play is represented by eighteen queues with precedence constraints, in series,
where the primitives are the random group playing times on each stage of a hole.
We characterize balanced courses and show the advantages over unbalanced
courses.
This paper presents a dynamic simulation model of patient’s health outcome and
length of stay based on initial health risk and physician’s assessment of risk.
Simulation results are empirically supported by analyzing a detailed dataset of
1,031 patients admitted to a large southeastern hospital in US.
4 - Rare-event Simulation for Vulnerability Analysis of Power Grids
John Shortle, George Mason University, 4400 University Dr.,
MSN 4A6, Fairfax, VA, United States of America,
jshortle@gmu.edu, Jie Xu, Chun-hung Chen
3 - Dynamic Personalized Monitoring and Treatment Control of
Glaucoma
Pooyan Kazemian,PhD Candidate, University of Michigan-Ann
Arbor, 1205 Beal Ave., Ann Arbor MI 48105, United States of
America, pooyan@umich.edu, Jonathan Helm, Mariel Lavieri,
Joshua Stein, Mark Van Oyen,
Vulnerability of a power grid can be evaluated by systematically considering
failures of individual elements and estimating the likelihood of a large-scale
blackout following these initial failures. This talk presents a method for
identifying vulnerable links by using a low-fidelity model of the power system to
guide simulation of a higher-fidelity model. Numerical examples using real power
systems are given.
We develop an innovative modeling framework for chronic disease patients to
help guide clinicians to quickly detect disease progression and adjust the
treatment plan over time to limit disease progression. The model is able to (1)
optimize the time interval between sequential monitoring tests; (2) specify the
best set of tests to take during each patient’s office visit; and (3) provide target
values for the controllable disease risk factors. Glaucoma is discussed as a case
study.
■ TD77
4 - Evaluating Consumer m-Health Services for Promoting Healthy
Eating: A Randomized Field Experiment
Yi-Chin Lin,CMU, 5000 Forbes Avenue, Pittsburgh PA 15213,
United States of America, yichinl@cmu.edu, Vibanshu Abhishek,
Julie Downs, Rema Padman
77-Room 300, CC
Supply Chain Optimization
Contributed Session
Chair: Robert Russell, Professor of Operations Management,
Univ. of Tulsa, 800 S Tucker Drive, College of Business, Tulsa, OK,
74104, United States of America, rrussell@utulsa.edu
1 - A Practical Application of Large-Scale Capacitated
Facility Location
Uday Rao, Professor, OBAIS Department, College of Business,
University of Cincinnati, Cincinnati, OH, 45221, United States of
America, uday.rao@uc.edu, Amit Raturi, Maria Caridi
Mobile apps have great potential to provide promising services to improve
consumers’ engagement and behaviors. Focusing on healthy eating, this study
shows that an image-based professional support greatly improves consumer
engagement and eating behaviors, while social media and a heuristic approach of
self-management might have negative effects in some occasions.
5 -Dynamic Matching in a Two-Sided Market
Yun Zhou,University of Toronto, 105 St. George Street, Toronto,
ON, Canada, Yun.Zhou13@Rotman.Utoronto.ca, Ming Hu
We study a large-scale capacitated facility location problem motivated by
interaction with a company in the US MidWest. The problem has 1000 locations
with demand seasonality, several facility types differing in capacity, fixed
installation and variable operating costs, and transportation costs depending on
transportation speed / mode selected. We present solution approaches using
mathematical programming, clustering and quick heuristics. We test sensitivity to
problem parameters.
A two-sided market often shares a common structure that engages three parties:
the supply side, the demand side and an intermediate firm facing intertemporal
uncertainty on both supply/demand sides. We propose a general framework of
dynamically matching supply with demand of heterogenous types (with
horizontally or vertically differentiated types as special cases) by the intermediary
firm and explore the optimal and heuristic matching policies.
372
2 - A Branch-and-Price Algorithm for Switchgrass Logistic Supply
Chain Design
Maichel M. Aguayo Bustos, Virginia Tech, 250 Durham Hall, 1145
Perry Street, MC 0118, Blacksburg, VA, 24061, United States of
America, maiaguay@vt.edu, Subhash C. Sarin, John S. Cundiff
TD79
2 - A Dynamic Garch Model for Energy Portfolio Allocation in
Electricity Markets
Reinaldo Garcia, Associate Professor, University of Brasilia - UnB,
Faculty of Technology, Industrial Engineering Department,
Brasilia, 70904-970, Brazil, rcgarcia@unb.br, Javier Contreras,
Virginia González, Janiele E. S. C. Custodio
Given the locations of a bio-energy plant and storage facilities for a switchgrassbased bio-ethanol supply chain, we introduce a multi-period mixed integer
programming model to determine both strategic and tactical decisions. A novel
branch-and price approach is used to obtain near-optimal solutions for large-sized
problem instances. Results of its implementation to a case study are also
presented.
In the deregulated electricity markets, a Generation Company (Genco) has to
optimally allocate their energy portfolio. Modern Portfolio Theory (MPT) allows a
Genco to maximize their profit and decrease their associated risk. This paper
proposes a model where MPT is combined with a Generalized Autoregressive
Conditional Heteroskedastic (GARCH) prediction model for a Genco to optimally
diversify their energy portfolio. The model is applied to the PJM electricity market
showing its capabilities.
3 - A Newsvendor Problem with Multiple, Capacitated Suppliers and
Marginal Quantity Discounts
Roshanak Mohammadivojdan, PhD Student, University of
Florida, 303 Weil Hall, P.O. Box 116595, Gainesville, FL, 32611,
United States of America, rmohammadivojdan@ufl.edu,
Joseph Geunes
3 - Intraday Electricity Load Forecasting using Rule-based Model
Myung Suk Kim, Professor, Sogang University, #1 Shinsu-Dong,
Mapo-Gu, Seoul, Korea, Republic of, myungkim@sogang.ac.kr
A rule-based model selection methodology incorporating a multiplicative seasonal
autoregressive with exogenous variables (ARX) model and a support vector
machine (SVM) is provided and applied to Korean hourly electricity load data. We
set up a rule that determines which of the SVM and ARX models should be
applied to forecasting a specific hour within a day. The proposed rule-based model
selection methodology outperforms its benchmarks.
We consider a newsvendor who may order stock from multiple, capacitated
suppliers, each of which offers a marginal quantity discount pricing structure. The
newsvendor seeks to minimize its total procurement plus expected overstock and
understock costs, resulting in an objective function consisting of a sum of convex
and concave terms. We provide an algorithmic approach that permits solving this
non-convex problem in pseudopolynomial time by solving a set of 0-1 knapsack
subproblems.
4 - Modeling Grid Operations in China’s Partially-Restructured
Electricity Market
Michael Davidson, Massachusetts Institute of Technology, 400
Main Street, E19-411, Cambridge, MA, 02139, United States of
America, michd@mit.edu, Valerie Karplus, Ignacio Perez Arriaga
4 - An Integrated Model for Supplier Selection and Optimal Order
Allocation Considering Uncertainty
Majid Hooshmandi Maher, Allameh Tabatabaei University,
Faculty of Management & Accounting, Hemmat Exp, Tehran,
Iran, majidhooshmand@gmail.com
Long transitions of restructuring vertically-integrated electric utilities can affect
interim market operations and assumptions underlying tools for policy
assessment. We develop a mixed integer unit commitment model of China’s
northeast region with several legacy central planning mechanisms modeled as
regulatory constraints and penalties. We analyze their influence on system
operation, test tractability of formulations and validate with actual operational
outcomes.
This paper first presents an approach for supplier evaluation based on integrated
multiple criteria decision making model and then proposes a mathematical model
to optimize the order allocation in a supply chain considering uncertainty in
different parameters. An Iranian automobile company is utilized as a case study.
The mathematical model is solved by genetic algorithm and the performance of
the model is verified using other optimization approaches.
5 - Fuel Hedging Strategy for Electric Power Utilities
Chung-Hsiao Wang, LG&E and KU, 102 Spruce Ln, Louisville,
KY, United States of America, chunghsiao@hotmail.com,
Kyung Jo Min
5 - The Production Routing Problem with Vehicle Costs
Robert Russell, Professor of Operations Management, Univ. of
Tulsa, 800 S Tucker Drive, College of Business, Tulsa, OK, 74104,
United States of America, rrussell@utulsa.edu
In recent years, natural gas combined cycle power plants have started to replace
aging and less efficient coal power plants. Because fuel costs represent the
majority of total costs for an electric power utility, how to manage volume and
price risks for coal and natural gas fuel is critically important. In this paper, we
develop mathematical models for structured and analytical guidelines on fuel
hedging strategies for a utility owning both types of generation units.
This paper addresses the integration of production, inventory, distribution, and
vehicle costs for supplying retail demand locations from a production facility. A
mixed integer model is used to determine an approximate solution to the
production routing problem with vehicle costs and a vehicle routing
metaheuristic is used to sequence routes for each time period. Computational
results are reported and compared to results from the traditional production
routing problem.
■ TD79
79-Room 302, CC
■ TD78
78-Room 301, CC
Invited Session
Electricity Markets and Utilities
Contributed Session
1 - SAS - Building and Solving Optimization Models with SAS
Ed Hughes, Principal Product Manager, SAS, Rob Pratt,
David Kraay
Chair: Chung-Hsiao Wang, LG&E and KU, 102 Spruce Ln, Louisville,
KY, United States of America, chunghsiao@hotmail.com
1 - Analysis of Consumer Behavior Towards Dynamic Residential
Electricity Pricing
Prajwal Khadgi, PhD Candidate, University of Louisville,
Speed School of Engineering, Louisville, KY, 40219,
United States of America, p0khad01@louisville.edu, Lihui Bai
SAS provides a broad spectrum of data and analytic capabilities, including
statistics, data and text mining, econometrics and forecasting, and operations
research-optimization, simulation, and scheduling. OPTMODEL from SAS
provides a powerful and intuitive algebraic optimization modeling language and
unified support for building and solving LP, MILP, QP, NLP, CLP, and networkoriented models. We’ll demonstrate OPTMODEL for basic and advanced
problems, highlighting its newer capabilities and its support for both standard and
customized solution approaches.
Variable electricity pricing for the control of residential load has attracted much
interest in the field of demand response, and static variable pricing such as time of
use rates has had successful applications in the US as an optional service.
However, dynamic variable pricing remains an open question, due to lack of
understanding on consumer behavior. We study consumer behavior against two
dynamic rates, i.e., demand charge and load-following rates, using utility
functions and simulation.
2 - Responsive Learning Technologies - Online Games to Teach
Operations and Supply Chain Management
Sam Wood, President, Responsive Learning Technologies
Learn about online competitive exercises that are used in Operations
Management courses and Supply Chain Management courses to teach topics like
capacity management, lead time management, inventory control, supply chain
design and logistics. These games are typically used as graded assignments
373

Tuesday - Conference Calendar

Transcription

Similar documents

For far too long, City Hall – and its signature courtyard

THEDROPBOXFILM.ca

Philadelphia Cream Cheese

PDR-13-30-10:Layout 1.qxd - Philadelphia Public Record

Professionals Gather for the Philadelphia Diversity and Leadership

Welcome Brochure

PDF Press Kit

Catholic Community Center

Tough Shit - Jack Fritscher

luxury apartments | bala cynwyd