2016 INFORMS Annual Meeting Program

190 / 561

INFORMS Nashville – 2016
190
MC24
109-MCC
Personalizing Healthcare Decision-Making
Sponsored: Health Applications
Sponsored Session
Chair: Anil Aswani, UC Berkeley, 1, San Francisco, CA,
United States,
1 - Incentive Design In The Medicare Shared Savings Program
Auyon Siddiq, UC Berkeley, Berkeley, CA, United States,
Anil Aswani, Zuo-Jun (Max) Shen
The Medicare Shared Savings Program (MSSP) was created to control rising
healthcare costs by offering Medicare providers financial incentives to reduce
spending on healthcare delivery. We place the MSSP within a principal-agent
framework and investigate the impact of directly subsidizing investments made by
a provider to improve operational efficiency. We then estimate parameters of the
principal-agent model using financial performance data of Medicare providers
participating in the MSSP. Our analytical and empirical results both suggest that a
direct subsidy can yield net reductions in total Medicare spending, despite the
additional payments made to providers.
2 - Approximation Algorithms For Population-scale Personal
Dietary Management
Pedro Hespanhol, UC Berkeley,
Anil Aswani
Diet is an important component of wellness and health; however, various fiscal,
geographic, and time constraints can make it difficult for families to healthfully
manage dietary decisions and purchases. This talk describes a novel mixed-integer
formulation for personal management of dietary decisions, and this formulation
can be generalized to a new class of knapsack-like problems. We design an
approximation algorithm to solve these problems, and such an approximation
algorithm enables scaling the use of our personal dietary management
formulation to the broader population.
3 - A Decision Analytics Approach For Clinical Intervention Design
Yonatan Mintz, UC Berkeley, Anil Aswani,
Philip Kaminsky, Yoshimi Fukuoka, Elena Flowers
When designing behavioral interventions it is crucial to take into account the
inherent tradeoff between quality of care and intervention cost. In this paper, we
develop an algorithm which uses patient data to resolve this tradeoff effectively in
the context of weight loss interventions. We expand on a previously developed
utility maximization model for patient behavior by utilizing integer programming
and Bayesian prediction to evaluate the efficacy of various weight loss
interventions and combine them into a weight loss program. We then present
simulation results which show that our method maintains efficacy while
potentially reducing the associated person hours and cost of the intervention.
4 - A Correlation-preserving Method To Estimate Risk Factor
Trajectories From Cross-sectional Data
Sze-chuan Suen, USC, Los Angeles, CA, United States,
Jeremy D. Goldhaber-Fiebert, Sanjay Basu
How should we approximate individual risk factors (i.e., cholesterol levels, BMI,
etc.) over time when we only have information about the distributions over the
whole population at each time period? We use a shortest-distance algorithm
which preserves correlation to approximate risk factor trajectories for use in
microsimulation models of disease when only cross-sectional data is available. We
compare the treatment implications of using this algorithm with other commonly
used methods.
MC25
110A-MCC
Project Related SCM I
Invited: Project Management and Scheduling
Invited Session
Chair: Xiaoqiang Cai, The Chinese University of Hong Kong, Shatin,
Hong Kong, Hong Kong,
1 - Capacity Control Policies For Leasing Industry Based On
Customers’ Behavior
Lifeng Zhang, University of Electronic Science and Technology of
China, Chengdu, China, , Yinping Mu,
Shiming Li
The paper studies the capacity control strategy for multiperiod and multiproduct
leasing based on customers’ behavior pricing strategy. Considering that the
customers’ behavior will affect the value of the products in the process of using
rental products, and then affect the benefit of the industry. We regard product
prices as a function of customers’ behavior, and build a model to analyze how to
solve the mismatching problem between capacity and demand with upgrades. We
present the stochastic dynamic programming formulation for customers’ behavior,
and propose a new product upgrade mechanism.Finally, we perform
computational expeiments to testify the qualities of the model.
2 - Downstream Firm’s Investment With Equity Holding In
Decentralized Assembly Systems
Hong Fu, University of Electronic Science and Technology of
China, Chengdu, China,
Yongkai Ma, Xiaoqiang Cai
We consider a decentralized assembly system in which $n$ upstream firms sell
complementary components to a downstream firm facing a stochastic and price-
sensitive demand. The downstream firm may make an investment to hold equity
in an upstream firm. This not only enables the downstream firm to share the
profit of the upstream firm as determined by the equity held, but also provides
the needed resources for the upstream firm to improve its production efficiency
and consequently benefits the entire system. We consider two distinct decision
settings: upstream Stackelberg and downstream Stackelberg. We characterize the
optimal decisions of the chain members, and obtain some useful insights.
3 - Ex-ante Transfer Pricing Decision In a Multinational Firm
Lianmin Zhang, Nanjing University, Nanjing, China,
Xiaopeng Zhang
This paper focus on the practice observed recently that MNFs do not only produce
for themselves but also for the local competitors. TP decision has ex-ante feature
in the analytical model and the arm’s length principle is applied also.
4 - Optimal Policies For Two-products Supply Chain With Free Gift
Cards Promotion
Yuefeng Li, University of Electronic Science and Technology of
China, Chengdu, China,
Jingming Pan, Xiaowo Tang
Many retailers offer free gift cards for attracting more consumers. These gift cards
are rewarded to consumers and can be redeemed on the purchase of other
products at the retailer. In this paper, we consider a supply chain system with two
independent manufacturers (M1 and M2) and one retailer. The retailer sells two
products, product 1 from the manufacturer 1 and product 2 from the
manufacturer 2. And she offers a specific “free” gift cards promotion, i.e., the
consumer who buy the product 1 can get a gift card then redeem gift card only to
buy product 2. Based on the above assumptions, we develop a decision model and
get the optimal strategies for the retailer.
MC26
110B-MCC
Optimal Auctions
Invited: Auctions
Invited Session
Chair: Ian Kash, Microsoft, Cambridge, United Kingdom,
1 - A Continuous Approximation Method For Optimal Auction Design
Eiichiro Kazumori, SUNY, Buffalo, NY, United States,
This paper propose a new method to analyze the optimal auction design problem.
The starting point is an observation that seller’s profit function is Baire class 1 that
can be derived as a pointwise limit of a sequence of continuous functions. Thus
the optimal auction mechanism is a limit of the nonlinear pricing problems. This
continuous approximation method can be regarded as an application of the path-
following method to the optimal auction design problem. Using this novel
method, we characterize the optimal auction mechanism with heterogeneous
objects and multidimensional types with continuous distributions by unifying
Myerson(1981), Mussa and Rosen(1978), and Rochet and Choné (1998).
2 - Strong Duality For A Multiple Good Monopolist
Christos Tzamos, MIT,
We provide a duality-based framework for revenue maximization in a multiple-
good monopoly. Our framework shows that every optimal mechanism has a
certificate of optimality, taking the form of an optimal transportation map
between measures. Using our framework, we characterize optimal mechanisms
showing that a mechanism is optimal if and only if stochastic dominance
conditions hold between specific measures induced by the buyer’s type
distribution. As a corollary, we consider the case of n independent uniform items
each supported on [c,c+1] and show that grand bundling is optimal if and only if
c is sufficiently large compared to n. This extends Pavlov’s result for 2 items
[Pavlov11].
MC24