2015 Informs Annual Meeting

WA52

INFORMS Philadelphia – 2015

WA51 51-Room 106B, CC Analyzing and Managing Incentives Sponsor: Manufacturing & Service Operations Management Sponsored Session Chair: Tinglong Dai, Assistant Professor, Johns Hopkins University, 100 International Dr, Baltimore, MD, 21202, United States of America, dai@jhu.edu Co-Chair: Fuqiang Zhang, Olin Business School, Washington University, St. Louis, MO, United States of America, fzhang22@wustl.edu 1 - Incorporating Customer Response in an Online-to-offline Fulfillment Strategy Elnaz Jalilipour Alishah, PhD Candidate, University of Washington, Seattle, Foster School of Business, Mackenzie Hall 358, Seattle, Wa, 98195-3200, United States of America, jalilipo@uw.edu, Yong-Pin Zhou We study an omni-channel retailer who fulfills online customer orders using inventory from offline retail stores. We examine the practice of fulfilling from the nearest location, and suggest other strategies in response to customer price and leadtime preferences. 2 - Signalling by Testing Tinglong Dai, Assistant Professor, Johns Hopkins University, 100 International Dr, Baltimore, MD, 21202, United States of America, dai@jhu.edu, Shubhranshu Singh Diagnostic experts, such as medical specialists, often only imperfectly observes customers’ true conditions, and may resort to advanced testing procedures. We model a diagnostic expert’s diagnostic decision tree problem when the expert’s level of competence is unknown to customers. We characterize perfect Bayesian equilibria, and prove the existence and uniqueness of a separating equilibrium. Our results offer insights into diagnostic experts’ testing behavior driven by their signaling efforts. 3 - Information Sharing in Competing Supply Chains with Production Cost Reduction Shilu Tong, Chinese University of Hong Kong, Shenzhen, 2001, Longxiang Blvd, Longgang District, Shenzhen, China, tongshilu@cuhk.edu.cn, Albert Ha, Quan Tian We consider the problem of sharing demand information in two competing supply chains, each consisting of one manufacturer and one retailer. A supply chain that engages in information sharing triggers a reaction from the rival chain. Such a reaction may benefit or hurt the first supply chain, depending on whether the retailers compete on quantity or price, and whether the manufacturers are efficient in cost reduction or not. 4 - Push vs. Pull: How to Best Allocate Supply Risk in Random Yield Supply Chains Guang Xiao, Olin Business School, Washington University in St. Louis, St. Louis, United States of America, xiaoguang@wustl.edu, Panos Kouvelis We consider a bilateral supply chain with supply random yield and propose three variants of wholesale price contracts, which induce different risk allocations between the supply chain parties. We completely characterize the Pareto set of any contract type combination to fully explore the price negotiation possibilities and profit improvement opportunities within the supply chain.

3 - Localized Assortment Selection Graham Poliner, Vice President Merchandise And Supply Chain Analytics, Macy’s Inc., 1440 Broadway, New York, NY, 10018, United States of America, graham.poliner@macys.com, Bhagyesh Phanse, Mark Posner We consider the assortment selection problem for omnichannel selling locations to support multiple fulfillment alternatives (e.g. direct to consumer, walk in store customers, and buy online pick up in store). The total assortment breadth offered is much larger than the capacity of a single store or fulfillment center. We describe how cross-channel demand information and product affinities can be used to inform localized assortment decisions, and we present the results of pilot implementations. 4 - Optimal Inventory Placement under Warehouse Capacity Constraints Salal Humair, Principal Research Scientist, Amazon.com, Inc., 500 Westlake Ave N, Seattle, WA, 98109, United States of America, salal@amazon.com, Onur Ozkok, Erdem Eskigun We consider how to place initial stocks for multiple products in multiple warehouses that serve multiple regions, when demand from a region can be fulfilled by more than one warehouse. We obtain the optimal initial stocks using a constructive algorithm with reasonable complexity. We use numerical experiments to understand how the algorithm distributes products across warehouses. We use these insights to decompose the problem of how much to buy vs. how to distribute it across the warehouses. WA50 50-Room 106A, CC New Models and Algorithms for Exploration and Exploitation Tradeoff Optimization Sponsor: Manufacturing & Service Operations Management Sponsored Session Chair: Retsef Levi, J. Spencer Standish (1945) Professor of Operations Management, Sloan School of Management, MIT, 100 Main Street, BDG E62-562, Cambridge, MA, 02142, United States of America, retsef@mit.edu 1 - Approximate Gittins Indices for The Multi-armed Bandits Multi-armed bandits is a fundamental problem that deals with the tradeoff between exploration and exploitation. While theoretical regret bounds have been proved for many solution methods, only in special cases have these bounds been shown to be optimal. By using Gittins’ theorem, that applies to the discounted infinite horizon MAB problem, we propose an efficient, asymptotically optimal algorithm, AGI, that is competitive with IDS, UCB and Thompson Sampling methods for Bernoulli rewards. 2 - Stochastic Selection Problems with Testing Chen Attias, Weizmann Institute of Science, 234 Herzl Street, We study a new class of stochastic selection problems, where the goal is to choose the minimal cost option among alternatives with a-priori stochastic costs, by introducing a testing option that has fixed cost but reveals the realization of the random cost of the option being tested. For several interesting cases of this model, we obtain the optimal testing order and show that local decision rules that only consider the value of the currently tested option are optimal. 3 - When and How to Test in Scheduling Nonhomogeneous Job Classes Yaron Shaposhnik, MIT, 77 Massachusetts Avenue, Bldg. E40- 149, Cambridge, MA, 02139, United States of America, shap@mit.edu, Retsef Levi, Thomas Magnanti We study a fundamental tradeoff that arises in many operational settings of performing work under uncertainty (exploitation) and investing capacity to reduce the underlying uncertainty (exploration). Sample domains include emergency departments, maintenance, scheduling, and project management. Our work leverages innovative analysis based on stochastic orders to obtain crisp descriptions of computationally efficient optimal exploration-exploitation policies, and provides managerial insights. Eli Gutin, Mr, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Cambridge, MA, 02139, United States of America, gutin@mit.edu, Vivek Farias Rehovot, 7610001, Israel, attiasc@mit.edu, Retsef Levi, Thomas Magnanti, Yaron Shaposhnik, Robert Krauthgamer

WA52 52-Room 107A, CC Retail Management I Contributed Session

Chair: Anurag Agarwal, Professor, University of South Florida, 8350 N Tamiami Tr, Sarasota, FL, 34243, United States of America, agarwala@sar.usf.edu 1 - Shrinking Shrinkage: An Empirical Investigation into Antecedents and Countermeasures Shivom Aggarwal, IE Business School, Instituto de Empresa, S.L., CIF: B823343, C/ Maria de Molina, 12 Bajo, Madrid, 28006, Spain, dr.shivom@gmail.com, Daniel Corsten Vendor-side fraud has been overlooked in literature due to intractability, but poses to be significant antecedent of shrinkage. Using multi-store longitudinal data from a US retailer, we investigate holistic antecedents of shrinkage and how they affect store performance. The unified framework will contribute to extant literature on efficient retail operations.

389