2016 INFORMS Annual Meeting Program

143 / 561

INFORMS Nashville – 2016
143
4 - Economic Effects Of European Union’s External Aviation Policy
Megersa Abate, Swedish National Road and Transport Research
Institute,
MA63
Cumberland 5- Omni
Continuous Space Location Modeling
Sponsored: Location Analysis
Sponsored Session
Chair: John Gunnar Carlsson, University of Southern California, 3715
McClintock Ave, Los Angeles, CA, 90089, United States,
1 - On The Dual And Rectangular Bounds For Continuous Facility
Location Problem
Nadere Mansouri, SMU,
Halit Uster
Lagrangian dual and Rectangular bounds (a rectangular distance location problem
specifically devised) are two of the lower bounding techniques for a continuous
facility location problem. We present results comparing these bounds at any
Weiszfeld iteration and upon convergence.
2 - Delivering Packages Jointly With A Truck And A Drone
John Gunnar Carlsson, University of Southern California,
One of the most talked-about developments in transportation and logistics in
recent years has been the potential use of drones for transporting packages. We
use a continuous approximation analysis to study a hybrid system in which
delivery trucks act as a mobile “base” for launching drones.
3 - The Competitive Facility Location Problem Under
Disruption Risks
Lawrence V Snyder, Associate Professor, Lehigh University,
200 West Packer Ave., Mohler Lab, Bethlehem, PA, 18015,
United States, , Ying Zhang, Ted K Ralphs
Two players sequentially locate facilities, competing to capture market share.
Facilities face disruption risks, and each customer seeks the nearest operational
facility for service, regardless of who operates it. The problem combines
competitive location and location with disruptions, an important combination
that has been absent from the literature. We model the problem as a Stackelberg
game, and formulate the leader’s decision problem as a binary bilevel
optimization problem. We propose a branch-and-cut algorithm and a variable
neighborhood decomposition search heuristic. Computational results suggest that
high quality solutions can be found quickly.
MA64
Cumberland 6- Omni
MCDA Methods and Applications
Sponsored: Multiple Criteria Decision Making
Sponsored Session
Chair: Roman Slowinski, Poznan University of Technology, Poland,
1 - Context Matters: Effects Of Product Type And Information
Overload On Choice Accuracy
Jyrki Wallenius, Professor, Aalto University School of Business,
Helsinki, Finland,
Pekka J Korhonen, Pekka Malo, Tommi Juhani Pajala,
Niklas Ravaja, Outi Somervuori
We report on the results of an experiment, which utilizes a new method for
generating many similar choice problems, enabling the objective measurement of
choice show that the product type matters for choice accuracy.
Moreover, we show that information overload is a relevant phenomenon in
MCDM experiments. However, what matters is the quality of information, not
just the quantity. When we add information that does not change the dominance
relations between products, choice accuracy is not degraded.
2 - Decision Under Risk And Uncertainty As A Multi-quantile
Decision Problem
Roman Slowinski, Poznan University of Technology, Poznan,
Poland,
Salvatore Corrente, Salvatorend,&nbGreco, Benedetto Matarazzo
We formulate the problem of decision under risk & uncertainty as a multiple
criteria decision problem, where criteria are some quantiles of the outcome
distribution, which are meaningful for the decision maker. To solve the multiple
criteria decision problem, we apply the robust ordinal regression approach. We
validate all the methodology on the classic newsvendor problem where we apply
GRIP and ELECTRE^GKMS methods to recommend a solution respecting
preferences of the decision maker.
MA65
Mockingbird 1- Omni
Gamification and User Engagement
Sponsored: Information Systems
Sponsored Session
Chair: Lei Wang, Pennsylvania State University, University Park, PA,
16801, AssigUnited States,
1 - Measuring The Impact Of Crowdsourcing On Mobile App User
Engagement And Retention: A Randomized Field Experiment
Zhuojun Gu, Pennsylvania State University,
Ravi Bapna, Jason Chan, AlokcustomGupta
In this paper, we propose a new strategy for enhancing mobile app user
engagement and retention by introducing crowdsourcing features that involve
users through the design of the app itself. We measure the causal impact of
crowdsourcing by conducting a randomized field experiment on a social mobile
game platform. We find higher user retention level could be achieved by allowing
users to submit content and customize their products. And sustained user
engagement and retention are enhanced most when both submission and access
options are available.
2 - Cultivate Consumer Engagement With Mobile And Gamification
Lei Wang, Pennsylvania State University, State College, PA, 16802,
United States, , Siyuan Liu
Despite the growing popularity of gamification and its potentials on customer
engagement, we still have very little knowledge about gamification and its impact
on customer engagement. In this research, we conduct a large-scale randomized
field experiment in a shopping mall in Asia to investigate the impact of
gamification on cultivating customer engagement. Our results will allow us to
effectively measure the causal impact of gamification and provide insights on
quantifying and improving the impacts of gamification on customer engagement
and mobile advertising. This study also provides important implications on how
firms could benefit from gamification.
3 - What Do Mobile Applications Bring a Longer Tail? An Empirical
Study Of Sales Concentration In Online Cchannels
Shahryar Doosti, University of Washington, Foster School of
Business, Mackenzie Hall, Seattle, WA, 98195, United States,
, Yong Tan, Youwei Wang
This work uses a dataset from a leading e-retailer which offers two online
channels, the desktop channel and mobile applications, to study the effect of long
tail on product sales in each channel. Our findings show that the long tail effect
exists in mobile application channel. In other words, there is more product
variation and less sales concentration on mobile app compared to desktop
channel.
MA66
Mockingbird 2- Omni
Model Calibration
Sponsored: Quality, Statistics and Reliability
Sponsored Session
Chair: Matthew Plumlee, University of Michigan, Ann Arbor, MI,
United States,
1 - Bayesian Calibration Of Inexact Computer Models
Matthew Plumlee, University of Michigan,
Bayesian calibration is used to study computer models in the presence of both a
calibration parameter and model bias. The parameter in the predominant
methodology is left undefined. Among other problems, this results in an issue
where the posterior of the parameter is sub-optimally broad. To date, there has
been no generally accepted alternatives. This paper proposes using Bayesian
calibration where the prior distribution on the bias is orthogonal to the gradient
of the computer model. Problems associated with Bayesian calibration are shown
to be mitigated through analytic results in addition to numerical and real
examples.
MA66