![Show Menu](styles/mobile-menu.png)
![Page Background](./../common/page-substrates/page0283.png)
INFORMS Nashville – 2016
281
2 - Human Interaction With Recommendation Systems:
On Bias And Exploration
Sven Schmit, Stanford University, Stanford, CA, United States,
schmit@stanford.edu, Ramesh Johari, Vijay Kamble,
Carlos Riquelme
More and more, recommendation systems assist users decision making. These
systems rely on historical data to provide suggestions. Little attention has been
paid to the interaction of humans and learning algorithms. We propose an explicit
model for the interaction between users and recommendations provided by a
platform. First, we show that this interaction leads to a bias in naive estimators
due to selection effects. This bias leads to suboptimal outcomes. Second, agents’
heterogeneous preferences lead to sufficient exploration of alternatives. Both
observations lead to new insights and practical advice.
3 - Uncertainty In Dynamic Matching With Application
To Organ Exchange
John Dickerson, Carnegie Mellon University, Pittsburgh, PA,
United States,
dickerson@cs.cmu.eduJohn Dickerson, University of Maryland, College Park, MD,
United States,
dickerson@cs.cmu.edu, Tuomas W Sandholm
We address dynamic kidney exchange, an organized market where patients with
end-stage renal failure swap willing but incompatible living donors. Specifically,
we focus on two types of uncertainty found in dynamic matching applications like
kidney exchange: probabilistic existence of edges in the current matching graph,
and determining the “best” way to match now given uncertainty over the future.
We discuss recent theoretical results and optimization methods to tackle this
uncertainty, and provide experimental results from applying these methods to
real-world data from the UNOS US-wide kidney exchange.
4 - Assortment Planning In School Choice
Peng Shi, Massachusetts Institute of Technology (MIT),
Cambridge, MA,
pengshi@mit.eduSchool choice systems assign students to schools by giving each student a menu of
options and eliciting their relative preferences among the options. Many such
systems uses the Gale-Shapely Deferred Acceptance (DA) algorithm, which also
take into account students’ priorities at each school. In previous literature, the
menus and priorities are considered to be given. In this paper, we study how the
school board can optimize the menus and priorities in order to maximum
students’ expected utilities subject to a budget for school busing. We apply this to
real data from Boston Public Schools.
TB52
214-MCC
Urban Transportation and Logistics in Public
Sector OR I
Sponsored: Public Sector OR
Sponsored Session
Chair: Sunghoon Chung, Binghamton University, The State University
of New York, P.O. Box 6000, Binghamton, NY, 13902, United States,
schung@binghamton.edu1 - A Vehicle Routing Problem For CNG Trucks With Fueling Stations
Yihuan (Ethan) Shao, University of Southern California,
Los Angeles, CA, United States,
yihuansh@usc.edu,
Maged M Dessouky
In this research, we consider a Vehicle Routing Problem (VRP) for Compressed
Natural Gas (CNG). Due to the limited number of available fueling stations and
small fuel tank capacity, CNG trucks call for special concerns with the refueling
problem. We introduce a new Mixed Integer Programming (MIP) model with
preprocessing and valid inequalities to solve the problem optimally, as well as a
new hybrid heuristic method to solve the problem in large scale. Numerical
experiments show the efficiency of our preprocessing and valid inequalities. We
also draw some insights from the experiments.
2 - The Truck Line Vehicle Routing Problem Of A Hub-and-spoke
Structure Based Parcel Carrier
Jaesang Park, Postech, Pohang, Korea, Republic of,
rose7@postech.ac.kr, Byung-In Kim, Ho-mahn Kwak,
Hyunjoon Kim, Jeongbin Kim
We handle a combined problem of trunk line vehicle routing and hub capacity
planning for a real-word parcel carrier, which uses a hub-and-spoke structure of
about 300 spokes and 11 hubs. Because of limited sorting capability, only ten
category groups can be classified at a spoke. The classified groups then are
delivered to an eligible hub, are sorted and delivered to appropriate spokes. It
handles about four million parcels daily and delivers them within two days using
4500 eleven tones trucks. A solution approach minimizing the total
transportation and hub operating cost is developed.
3 - Dynamic Bus Routing For Evacuation
Xiaohang Zhu, University at Buffalo, 2514 Deer Lakes Dr,
Amherst, NY, 14228, United States,
xzhu8@buffalo.edu,
Jee Eun Kang
Bus-based evacuation serves carless population and helps reduce congestion
during evacuations. This research develops a single level dynamic bus routing
problem, that determines the optimal bus routing with respect to dynamics of
traffic and congestion, that minimizes the total evacuation time experienced by all
evacuees. The proposed model is built upon the linear programming formulation
of the Cell Transmission Model (CTM).
TB53
Music Row 1- Omni
Innovations and Value Chain Management
Sponsored: Technology, Innovation Management &
Entrepreneurship
Sponsored Session
Chair: Ying-Ju Chen, HKUST Business School, Dept. of ISOM, HKUST,
Clear Water Bay, Kowloon, Hong Kong, 999077, Hong Kong,
imchen@ust.hkCo-Chair: Shihong Xiao, HKUST, Dept. of IELM, HKUST, Clear Water
Bay, Kowloon, Hong Kong, 999077, Hong Kong,
sxiaoab@connect.ust.hk1 - Impact Of Forecasting Accuracy On Sharing Forecasts In
Supply Chains
Hyoduk Shin, University of California - San Diego,
hdshin@ucsd.eduWe investigate the role of accuracy information of demand forecasts on sharing
forecasts within the supply chain between the downstream retailer and the
upstream supplier. We demonstrate how the uncertainty on the forecast accuracy
can help or hurt the ability to share demand forecasts.
2 - Supply Chain Contracts With Boundedly Rational Retailers
Behrooz Pourghannad, University of Minnesota, Minneapolis, MN,
United States,
behrooz.pourghannad@gmail.com,Guangwen
Kong, Tony H Cui
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.
3 - Milking The Quality Test: Improving The Milk Supply Chain Under
Competing Collection Intermediaries
Liying Mu, University of Delaware,
muliying@udel.edu,Milind Dawande, Xianjun Geng, Vijay Mookerjee
We examine operational and incentive issues that conspire to reduce the quality
of milk — via deliberate adulteration by milk farmers — under competing
collection stations in developing countries. Two recommendations are provided to
reduce the milk adulteration using minimal testing. Both solutions achieve a
socially-beneficial equilibrium outcome: All the farmers provide high-quality milk
and each competing station only conducts one mixed test and no individual
testing.
4 - 3d Printing Vs. Traditional Flexible Technology: Implication On
Manufacturing Strategies
Duo Shi, Washington University in Saint Louis, KH 401,
Olin Business School, 1 Brookings Drive, St. Louis, MO, 63130,
United States,
dshi@wustl.edu, Lingxiu Dong, Fuqiang Zhang
In this paper, we study a firm’s manufacturing strategies under two types of
flexible technologies: traditional flexible technology and 3D printing. The firm
adopts dedicated technology and one type of flexible technology, either the
traditional one or 3D printing. It has to choose an assortment from a potential set
of variants, assigns each variant to a production technology, and finally invests in
resource capacities. We find that traditional flexible technology and 3D printing
can have distinct technology assignment structures. Contrary to common
wisdom, the adoption of traditional flexible technology can reduce product
variety. 3D printing, however, always enhances product variety.
TB53