INFORMS Nashville – 2016
216
2 - Quantitative Approaches For Measuring And Relieving
Congestion In A Dialysis Unit At A Hospital: The Case Of
Compassionate Dialysis
Olga Bountali, Southern Methodist University, 3145 Dyer street,
Dallas, TX, United States,
obountali@smu.edu, Sila Çetinkaya,
Vishal Ahuja
We examine a congested healthcare delivery setting due to emergency treatment
of a chronic disease. A prominent example of the problem of interest is
congestion in the emergency room due to arrival of unfunded ESRD patients
needing dialysis (a.k.a., compassionate dialysis). Unfortunately, this is the only
treatment option for undocumented immigrants with ESRD., i.e., when their
clinical condition is deemed as life-threatening. We model the problem as a
queueing network with recurrent arrivals and service and examine various
performance metrics impacting patient welfare and social cost.
3 - Efficient Sampling From Large Databases For Controlled Trials
Sanjay Mehrotra, Professor, Northwestern University, Evanston,
IL, 60208, United States,
mehrotra@northwestern.edu,
Daniel Apley, Liwen Ouyang
Controlled trials are ubiquitously used to investigate the effect of a medical
treatment. The trial outcome can be dependent on a set of patient covariates.
Traditional approaches have relied primarily on randomized patient sampling and
allocation to treatment and control groups. We will present an approach that uses
optimal design of experiments (DOE) concepts to select the patients for the
treatment and control groups, based on their covariate values, in a manner that
optimizes the information content in the data.
4 - Optimal Stopping For Response-guided Dosing
Jakob Kotas, PhD, University of Portland, Portland, OR, 37221,
United States,
kotas@up.eduOne important decision in response-guided dosing is when to stop treatment. In
practice, stopping often occurs for patients in remission, or due to a finding of
futility or a desire to switch to a different drug. We have previously developed a
stochastic dynamic programming framework for response-guided dosing which
tailors dose to a patient’s stochastic evolution of disease state. In this talk we
present an extension of that model that allows for stopping before the scheduled
end time. Our numerical simulations show optimality of stopping below a
threshold state. We also outline structural properties of the stopping formulation,
including monotonicity of the threshold state with respect to time.
MD25
110A-MCC
Project Related SCM II
Invited: Project Management and Scheduling
Invited Session
Chair: Xiaoqiang Cai, The Chinese University of Hong Kong, Shatin NT,
xqcai@se.cuhk.edu.hk1 - Procurement Strategies With Asymmetric Yield Information
Houcai Shen, Nanjing University,
hcshen@nju.edu.cnIn this paper, we study firm purchase behavior when her supplier has private
information about his yield rate.
2 - Optimal Design Of Omnichannel
Xiaolin Xu, Nanjing University,
xuxl@nju.edu.cnNowadays omnichannel has received much attention in retail industry. However,
how to design an appropriate omnichannel strategy is a challenge facing the
retailers. For example, retailer can provide BOPS (buy online and pickup at store)
or STOP (shipment to store for consumers to pickup at store), in addtion to the
traditional O2O (online vs. offline) channels. The channel BOPS shares the store
inventory with the offline channel, while the channel STOP shares the DC
(distribution center) inventory with the online channel. We investigate how these
two omnichannel strateties affect the market segmentation, fill rates at store and
the retailer’s profit.
3 - Pricing And Ordering Strategies Of Supply Chain With Selling
Gift Cards
Wenqing Shi, University of Electronic Science and Technology of
China, Chengdu, China,
shiwenqing1218@163.com,
Jingming Pan, Xiaowo Tang
Gift cards is one popular promotion way in market, which is widely used to
replace traditional gift cash and gift products, especially when gift givers do not
know gift receivers’ performances. Basing on this phenomenon, we develop a
Stackelberg model to analyze the supplier’s and retailer’s optimal strategies with
selling gift cards. We also exam the influences of different parameters on the
optimal decisions and the supply chain performance.
4 - Product Quality Choice In Two-echelon Supply Chains Under
Post-sale Liability: Insights From Wholesale Price Contracts
Jianchang Fan, Doctor, University of Electronic Science and
Technology of China, Qingshuihe Campus: No.2006,
Xiyuan Ave, West Hi-Tech Zone, Chengdu, 611731, China,
love.andyfan@163.com,Debing Ni
A two-stage game model in a supply chain consisting of a manufacturer (M) and
a Retailer (R) is built via a wholesale price contract. M is liable for the harm
caused by its low-quality products. In the equilibrium, we find that (1) in spite
that product liability positively affects the wholesale price, M’s quality level, the
contracted quantity and the supply chain members’ profitability are independent
of it; (2) in response to changes in liability-related factors, the product quality is in
conflict with the profits for both M and R, but the quality and the profit increase
in quality improvement efficiency; (3) the wholesale price serves as a medium for
M to share its ex ante expected liability cost with R.
MD26
110B-MCC
Auctions, Markets, and Mechanism Design
Invited: Auctions
Invited Session
Chair: Brian Baisa, Amherst College, Amherst, MA, United States,
bbaisa@amherst.edu1 - Mechanism Design For Team Formation
Greg Leo, Vanderbilt University, Nashville, TN, United States,
g.leo@vanderbilt.edu, Martin van der Linden, Jian Lou,
Yevgeniy Vorobeychik, Myrna Wonders
Team, or coalition, formation is a fundamental problem in AI and game theory.
While there has been extensive research on coalitional stability, especially in
hedonic games, the problem of designing mechanisms for coalition formation has
received relatively little attention. Theoretically, most of what is known about
such design problems is negative; for example, no mechanism that always
matches soulmates (individuals who prefer to be matched together) is
strategyproof in general. We describe some of our recent positive results about
achieving efficiency, incentive compatibility, individual rationality, and coalitional
stability in important restricted classes of hedonic preferences.
2 - Equilibrium In a Uniform-Price Auction withPrivate Values
Justin Burkett, Wake Forest University,
burketje@wfu.eduIn a model incorporating bidders with private information bidding for multiple
units of a homogeneous good, I show that the equilibrium bid functions can be
completely characterized, sometimes in closed form. The equilibrium bid
functions can be considered aggregated bid functions from first-price single unit
bidders. This property allows one to extend results from the single-unit auction
literature to this multi-unit setting.
3 - Strategic Ironing In Pay-as-bid Auctions: Equilibrium Existence
With Private Information
Kyle Woodward, University of North Carolina at Chapel Hill,
Gardner Hall, CB 3305, Chapel Hill, NC, 27599-3305,
United States,
kyle.woodward@unc.eduI establish the existence of pure-strategy Bayesian-Nash equilibria in divisible-
good pay-as-bid auctions with private information, and show that such equilibria
can approximate equilibria in nearby multi-unit auctions. I show that equilibrium
strategies exhibit strategic ironing, a reduction of bids below what might be
expected. Strategic ironing has implications for the tractability of equilibrium
strategies.
4 - Efficient Multi-unit Auction Design Without
Quasilinear Preferences
Brian Baisa, Amherst College,
bbaisa@amherst.eduI study efficient multi-unit auction design when bidders have private values,
multi-unit demands, and non-quasilinear preferences. Without quasilinearity, the
Vickrey auction loses it desired incentive and efficiency properties. I construct a
novel mechanism that retains the desirable properties of Vickrey auction, even in
cases when bidders have non-quasilinear preferences. When bidders have single
dimensional types, the mechanism (1) is dominant strategy incentive compatible,
(2) is Pareto efficient, and (3) provides no subsidies. If bidders types are multi-
dimensional, I show that there is no mechanism that satisfies these three
properties.
MD25