2016 INFORMS Annual Meeting Program

SC37

INFORMS Nashville – 2016

3 - Is The FDA Too Conservative Or Too Aggressive?: A Bayesian Decision Analysis Of Clinical Trial Design

We analyzed 30 months of a retailer’s history on store-month sales and potential sales drivers to measure the impact of store selling staff level on revenue. We identified a third of the stores where our analysis indicated that increasing staffing would increase sales. The retailer confirmed this finding via a 16 store test which showed that a 10% increase in sales staff resulted in a 9.9% sales increase, and was highly profitable. The retailer is now implementing our finding in other stores. SC36 205B-MCC Information and Risk in Supply Chain Networks General Session Chair: Kostas Bimpikis, Stanford Graduate School of Business, 650 Knight Way, Stanford, CA, 94305, United States, kostasb@stanford.edu 1 - Inventory Management With Censored Demand Data: The Adversarial Case Michail Markakis, Universitat Pompeu Fabra, mihalis.markakis@upf.edu We consider a repeated newsvendor problem where the demand distribution is unknown ex ante and has to be learned from sales/censored data. To shed light to scenarios where the demand may be non-stationary, e.g., exhibiting trends or seasonalities, we model the problem as a game between the inventory manager and an oblivious opponent, who prior to the game decides a sequence of demands for the different periods arbitrarily. We propose randomized inventory management policies that perform well with respect to the regret criterion, i.e., the difference between a policy’s cumulative cost and the cumulative cost of the best fixed action/ordering decision in hindsight, for any given demand sequence. 2 - Optimizing Local Content Requirements Under Technology Gaps Shiliang Cui, Georgetown University, shiliang.cui@georgetown.edu, Lauren Xiaoyuan Lu We study the optimal Local Content Requirements (LCR) and innovation policies of a developing economy in which a foreign Original Equipment Manufacturer (OEM) produces and sells a final product. We find that as the domestic component supply base becomes more cost efficient, surprisingly, the OEM’s profit could decrease. 3 - Take-rate Crowdsourcing Contracts Yun Zhou, University of Toronto, Toronto, ON, Canada, yzhou.zj@gmail.com, Ming Hu Motivated by the surge pricing strategy by the ridesharing platforms, we consider the pricing problem in a two-sided market. The total amount of supply is an increasing function of the wage and the amount of demand depends on the price. We model supply and demand uncertainty by a number of different scenarios, and show that the take-rate price contract is optimal for maximizing the platform’s profit or the total utility of the platform and the supply side when only the market size is scenario dependent. In more general cases, we derive performance bounds for the take-rate contract. SC37 205C-MCC Revenue Management, Assortments and Choice Models Sponsored: Manufacturing & Service Oper Mgmt, Supply Chain Sponsored Session Chair: Ozge Sahin, Johns Hopkins University, Brooklyn, NY, United States, ozge.sahin@jhu.edu 1 - Consumer Choice Under Rational Inattention And Implications For Assortment Planning We study the choice behavior of rationally inattentive customers who optimally acquire information about available options with ex-ante uncertain values through potentially different channels with different costs. Customers trade-off the benefits of better information obtained by asking questions with the associated cost. We quantify acquired information and its cost through a novel function based on conditional mutual information. We solve the consumer’s choice problem and analytically characterize the resulting optimal choice behavior. We illustrate some properties of the choice behavior and discuss implications for assortment planning. Tamer Boyaci, ESMT Berlin, Berlin, 10178, Germany, Tamer.Boyaci@esmt.org, Frank Huettner, Yalcin Akcay

Andrew W Lo, Charles E. and Susan T. Harris Professor, MIT, 100 Main Street, E62-618, Cambridge, MA, 02142, United States, alo-admin@mit.edu, Andrew W Lo, Charles E. and Susan T. Harris Professor, MIT, 32 Vassar Street, Cambridge, MA, 02139, United States, alo-admin@mit.edu, Leah Isakov, Vahid Montazerhodjat We explore the application of Bayesian decision analysis (BDA) to minimize the expected cost of drug approval, where the relative costs of Type I and Type II errors are calibrated using burden of disease data. For terminal illnesses with no existing therapies such as pancreatic cancer, the standard Type I error threshold of 2.5% is substantially more conservative than the BDA-optimal threshold of 23.9% to 27.8%. We compute BDA-optimal sizes for 25 of the most lethal diseases and show how a BDA-informed approval process can incorporate all stakeholders’ views in a systematic, transparent, internally consistent, and repeatable manner. 4 - A Comparison Between The Robust Risk-aware And Risk-seeking Managers In R&D Portfolio Management Aurelie Thiele, Associate Professor, Southern Methodist University, Dallas, TX, United States, aurelie@alum.mit.edu Shuyi Wang We analyze via simulation two mathematical modeling frameworks that reflect different managerial attitudes toward upside risk in R&D portfolio selection. The manager seeks to allocate a development budget between low-risk, low-reward projects, called incremental projects, and high-risk, high-reward projects, called innovational projects. We study the differences in strategy and portfolio’s risk profile that arise between a risk-aware manager, who takes upside risk because he has to for the long-term competitive advantage of his company, and a risk-seeking manager, who will take as big a bet as allowed by the model. SC35 205A-MCC Frontiers of Supply Chain Research Sponsored: Manufacturing & Service Oper Mgmt, Supply Chain Sponsored Session Chair: Karen Zheng, MIT Sloan School of Management, Cambridge, MA, 02142, yanchong@mit.edu 1 - Dual Co-product Technologies: Implications For Process Development And Adoption Brian Tomlin, Tuck School of Business, brian.tomlin@tuck.dartmouth.edu, Ying-Ju Chen, Yimin Wang Many industries operate technologies in which multiple outputs (co-products) are jointly produced. Three important attributes of a co-product technology are its production cost, its overall yield, and its co-product split. Process development often wrestles with an inherent trade-off: improvement in one attribute comes at the expense of another. In this talk, we first explore production and pricing decisions for a firm with two technologies and then use this foundation to examine implications for process development and process adoption. 2 - Impact Of Grocery Store Density And Market Structure On Food Waste Elena Belavina, University of Chicago Booth School of Business, elena.belavina@chicagobooth.edu Food waste is one of the major contributors of greenhouse gas emissions. If food waste was a country, it would be third largest polluter shortly after US and China. About $1 trillion dollars of food is wasted every year, which is equivalent to 1% of GDP globally. This study explores the impact of store density and market structure on consumer food waste. 3 - Self-policing In A Supply Chain Under Threat Of Public Disclosure Sang-Hyun Kim, Associate Professor, Yale University, New Haven, CT, United States, sang.kim@yale.edu, Saed Alizamir We study incentive dynamics among supply chain members and an external stakeholder (e.g., NGO) that impact environmental performance. A buyer inspects a supplier’s production in its supply chain to detect and correct environmental compliance violations. The buyer’s primary motive is to deter the NGO from discovering the violation first and publicize it, from which the buyer incurs a reputational penalty. The buyer and the NGO engage in a game to competitively set their inspection intensities, which influence the supplier’s decision to restore compliance. Together, the actions made by all parties determine the environmental outcome and and social welfare. 4 - Increasing Retail Sales Via Improved Store Staffing: An Empirical Study With Implemented Results Santiago Gallino, Dartmouth College, santiago.gallino@tuck.dartmouth.edu, Marshall L Fisher, Serguei Netessine

81