2016 INFORMS Annual Meeting Program

SC38

INFORMS Nashville – 2016

2 - Assortment Optimization For Choosy Customers Jake Feldman, Washington University in St. Louis, jbfeldman@wustl.edu

We identify the key drivers, market heterogeneity and production cost, for the controversy and provide managerial and political implications. Interestingly, the innovator can be better off with a strong competitor when market inequality is low or the competitor is strong enough.

We study two different choice models that capture the purchasing behavior of customers who only consider purchasing one of two substitutable products. We refer to these customers as choosy. The first choice model captures substitution behavior through probabilistic transitions between products. The second choice model that we study assumes each customer is characterized by a ranking of the products. An arriving customer will purchase her highest ranked product that is offered. Since we model choosy customers, we assume that these rankings contain at most two products. This paper focuses on the assortment optimization problem under these two choice models. 3 - Learning Consumer Tastes From Dynamic Assortments: A Nonparametric Bayesian Model Canan Ulu, Georgetown University, cu50@georgetown.edu Dorothee Honhon We study dynamic assortment decisions of a firm learning about consumer tastes by observing sales. Each period, the firm offers an assortment to maximize expected total profits over a finite horizon given its beliefs on consumer tastes. The consumers then choose a product that maximizes their utility and the firm updates its beliefs on consumer tastes after having observed sales. We model consumer tastes as locations on a Hotelling line and develop a nonparametric Bayesian learning model using Polya tree priors. We develop upper bounds on the firm’s total profit based on information relaxations and study the performance of various heuristic policies with respect to these upper bounds. 4 - The Price Of Flexibility Hoda Bidkhori, Swanson School of Engineering, University of Pittsburgh, Pittsburgh, PA, United States, bidkhori@pitt.edu Dimitris Bertsimas, Albert Dunning Process flexibility is a popular operations strategy that has been employed in many industries to help firms respond to uncertainty in product demand. Additional flexibility comes at a cost that firms must balance against the reduction of risk it can provide. We reduce the price of flexibility by taking an optimization approach to the process flexibility design problem. Unlike many approaches in the literature, we consider systems that may have nonhomogenous parameters and unbalanced capacity and demand. We formulate the problem as a robust adaptive optimization model, and propose a computationally tractable method for solving this model using standard integer optimization software. SC38 206A-MCC Collaborative New Product Development Sponsored: New Product Development Sponsored Session Chair: Niyazi Taneri, Singapore University of Technology and Design, Singapore, Singapore, na, Singapore, niyazitaneri@sutd.edu.sg 1 - The Role Of Operations In Alliances For New Product Development Niyazi Taneri, Singapore University of Technology and Design, niyazitaneri@sutd.edu.sg, Arnoud De Meyer We review contract theory and hypothesize its implications for the choice between collaborative alliances (where both parties exert joint efforts) and sequential alliances (where, for the most part, the partner takes over going forward). We test these hypotheses through the analysis of a dataset of over 2000 biopharmaceutical alliances. 2 - Optimal Sequential Investments In Product Development With Exogenous Technologies And Learning Shantanu Bhattacharya, Singapore Management University, shantanub@smu.edu.sg, Stylianos Kavadias, Sameer Hasija We determine the optimal investments for a firm when the product development opportunities come over time from two distinct exogenous technologies. Upfront investment in a product platform from a technology that is currently available gives higher returns from opportunities based on the platform technology in the future, due to the associated learning effects. We formulate the resource allocation problem and characterize the optimal development investments that determine the firm product development roadmap. We show that the firm’s optimal resource investment in platform development has a nuanced relationship with the relative speed of arrival of the new technology. 3 - Business Model For Technology-intensive Supply Chains Junghee Lee, University of California, San Diego,

SC39 207A-MCC Market Microstructure and Optimal Trading Sponsored: Applied Probability Sponsored Session Chair: Costis Maglaras, Columbia University, New York, NY, United States, c.maglaras@columbia.edu Co-Chair: Ciamac Cyrus Moallemi, Columbia University, New York, NY, United States, ciamac@gsb.columbia.edu 1 - Trading The Close — Market Impact And Optimal Execution Costis Maglaras, Columbia University, c.maglaras@columbia.edu The “close” concentrates a significant amount of daily liquidity for various market structure reasons. In this talk I will describe a market impact model for “Market- On-Close” (MOC) orders, and its consequences on optimal execution profiles. 2 - Portfolio Liquidity Estimation And Optimal Execution Kai Yuan, Columbia University, kyuan17@gsb.columbia.edu We develop a tractable model to estimate portfolio liquidity costs through a multi- dimensional generalization of the optimal execution model of Almgren and Chriss. Our model allows for the trading of standardized liquid bundles of assets (e.g., ETFs or indices). We show that in a “large universe” asymptotic limit, where the correlations across a large number of assets arise from relatively few underlying common factors, the liquidity cost of a portfolio is essentially driven by its idiosyncratic risk. Moreover, the additional benefit of trading standardized bundles is roughly equivalent to increasing the liquidity of individual assets. 3 - Optimal Execution In Hong Kong Given A Market-on-close Benchmark Christoph Frei, University of Alberta, Edmonton, AB, Canada, cfrei@ualberta.ca, Nicholas Westray For stocks traded on the Hong Kong Exchange, the median of five prices taken over the last minute of trading is currently chosen as the closing price. We introduce a stochastic control formulation to target such a median benchmark in an empirically justified model which takes the key microstructural features into account. We solve this problem by providing an explicit and efficient algorithm which can be used for the dynamic linear approximation of any square-integrable random variable. Implementing the algorithm on the stocks of the Hang Seng Index, we find an average improvement of around 6% in standard deviation of slippage compared to an average trader’s execution. 4 - Mean Field Games Of Singular Control With Applications Joon Seok Lee, UC Berkeley, 2938 McClure Street, # A207B, Oakland, CA, 94609, United States, ljshope@berkeley.edu Xin Guo We introduce a mean field game framework with singular controls. To solve this singular control problem with multiple agents, we derive the Kolmogorov forward equation for the singular control, which is a generalization of the mean fi eld game with regular controls. Both single controls with a bounded velocity and singular controls with a finite variation will be analyzed. Finally, we will present some applications to real options and systemic risk.

SC40 207B-MCC Computational Issues in Productivity and Efficiency Measurement Invited: Data Envelopment Analysis Invited Session

Chair: Jose Dula, Virginia Commonwealth University, Snead Hall, 301 W. Main Street, Richmond, VA, 23284, United States, jdula@vcu.edu 1 - Validating DEA As A Rating Tool: The Case Of CMS’s Nursing Home Compare. Jose Dula, Virginia Commonwealth University, jdula@vcu.edu Marie-Laure Bougnol The US government’s CMS agency rates more than 15000 nursing homes nationwide using a star system. The outcomes are disseminated in various ways including a user friendly and informative web page. We report on a study comparing the government’s ratings with classifications obtained with DEA using the same model and data. We answer the question: How would DEA fare as a tool to rate complex entities such as nursing homes?

junghee.lee@rady.ucsd.edu Krishnan Vish, Hyoduk Shin

In technology licensing, controversy has swirled among firms and policymakers about royalty base choice between subsystem and full system. We study the impact of royalty base on innovator’s business model decisions from R&D investment to manufacturing integration in Technology Intensive Supply Chain.

82