Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MD12

4 - Impact of Information Asymmetry and Limited Production Capacity on Business Interruption Insurance Yuan-Mao Kao, Duke University, Durham, NC, 27708, United States, N. Bora Keskin, Kevin Shang We study adverse selection and moral hazard issues that arise when an insurer offers business interruption insurance to a firm for guarding against disruption risks. The insurer cannot observe the firm’s demand forecasts and recovery effort when a disruption occurs. We characterize the optimal insurance contracts to deal with the information asymmetry, and show how the firm’s limited production capacity impacts the insurer’s contract design. We also analyze the impact of ignoring the information asymmetry in designing insurance contracts.

n MD13 North Bldg 126B Behavioral, Incentive and Policy Issues in Healthcare Sponsored: Manufacturing & Service Oper Mgmt/Healthcare Operations Sponsored Session Chair: Tinglong Dai, Johns Hopkins University, Johns Hopkins University, Baltimore, MD, 21202, United States Co-Chair: Ozlem Yildiz, University of Virginia, Charlottesville, VA, United States 1 - Physician Peer Effects on Speed and Quality: Evidence from the Emergency Department Raha Imanirad, Harvard Business School, Boston, MA, 02472, United States, Soroush Saghafian, Stephen J. Traub In this study, we estimate peer effects in the context of an Emergency Department (ED) setting by addressing the question of whether peer physicians’ characteristics including relative performance, experience, type of medical degree, and gender affect a physician’s performance. Our findings provide strong evidence on the existence of peer effects in this setting. Our results have important practical implications for improving the operations of EDs. These include superior physician scheduling where one needs to decide which providers should be scheduled during the same shift as well as in physician training where one needs to provide guidance to physicians with ways to improve their performance. 2 - Can Yardstick Competition Reduce Waiting Times? Ozlem Yildiz, University of Virginia, Darden School of Business, 100 Darden Blvd. Darden School of Business., Office: FOB 189, Charlottesville, VA, 22903, United States, Nicos Savva, Tolga Tezcan In this paper, we first show that the hospital reimbursement system currently used in practice does not incentivize hospitals to reduce waiting times. We then propose a modification which can achieve socially optimum investment without placing an onerous informational burden on hospital payers. 3 - Conditions of Participation: Inducing Organ Discards and Patient Deaths on Transplant Waiting Lists? Mohammad Delasay, Assistant Professor, Stony Brook University, Stony Brook, NY, United States, Sridhar R. Tayur We investigate waiting list management from the viewpoint of a transplant center in the presence of the conditions of participation (COP). COP penalizes centers with lower-than-expected post-transplant survivals. We model the waiting list as a multi-dimensional queue, which we analyze by proposing accurate truncation bounds of its state space. Our experiments illustrate that a single COP threshold setting scheme across diverse transplant centers without consideration of transplant centers’ specifications could induce cherry-picking of the candidates and offered organs. We discuss that COP lacks a net benefit approach by focusing only on post-transplant outcomes. 4 - Optimal Sequencing of Diagnostic Tests for Differential Diagnosis Simrita Singh, Northwestern University, Evanston, IL, United States, Sarang Deo, Sumit Kunnumkal Motivated by the challenges of diagnosing TB from other respiratory illnesses, we consider the problem of optimal differential diagnosis: The decision maker has to identify the correct underlying condition in the patient by sequentially conducting imperfect diagnostic tests, each aimed at diagnosing a single condition. We model this problem as a stochastic dynamic program and obtain partial characterization of the optimal policy. We employ a novel data-driven implementation of information relaxation approach to construct a computationally efficient and tight upper bound. We also develop a policy improvement algorithm based on a greedy heuristic that yields near-optimal performance.

n MD12 North Bldg 126A

Joint Session MSOM/Practice Curated: Choice Modeling and Applications in Retail Operations Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Aydin Alptekinoglu, Pennsylvania State University, Pennsylvania State University, University Park, PA, 16802, United States 1 - Product Line Design under Multinomial Logit Choices Hongmin Li, Arizona State University, WP Carey School of Business, Dept of Supply Chain Management, Tempe, AZ, 85287, United States, Scott Webster, Gwangjae Yu We study a product-line design problem in which customer choice among multiple products is given by a multinomial logit (MNL) model. A firm determines product attributes and prices in an evolving product line to maximize profit. In particular, given the prices and attributes of products that already exist in a product line, the firm optimizes prices and/or attributes of the new products to be added to the same product line. 2 - Dynamic Pricing for Varying Assortments Emily Mower, Harvard University, Cambridge, MA, United States, Kris Johnson Ferreira Most multi-product demand learning and dynamic pricing algorithms learn product-specific demand parameters as opposed to attribute-specific demand parameters. We develop an attribute-specific learning-then-earning dynamic pricing algorithm geared for companies whose assortments change over time. To maximize efficiency in the learning phase, we incorporate methods from conjoint analysis and optimal experimental design. We test our algorithm in a randomized controlled field experiment at an e-commerce platform that sells excess inventory John H. Semple, Southern Methodist University, Cox School of Business, 6212 Bishop Boulevard, Dallas, TX, 75205, United States, Aydin Alptekinoglu We investigate the problem of purchasing a bundle of different (but substitutable) products for future consumption. We term this bundle an “n-pack. The optimal consumption of the pack can be analyzed using dynamic programming, which we use to derive the optimal policy and the structure of the value function for this multi-state problem. In some cases, the value function can be given in closed form, and thus our problem does not suffer from the usual curse of dimensionality. 4 - A Comparative Empirical Study of Discrete Choice Models in Retail Operations Gustavo J. Vulcano, Universidad Torcuato di Tella, Av Figueroa Alcorta 7350, Suite 405, Buenos Aires, 1428, Argentina, Gerardo Berbeglia, Agustin Garassino In this paper, we conduct a systematic, empirical study of different demand models and estimation algorithms, spanning both maximum likelihood and least squares criteria. Through an exhaustive set of numerical experiments on synthetic and real data, we provide comparative statistics of the quality of different choice models and estimation methods, and characterize operational environments suitable for different model/estimation implementations. of fitness studio classes offered the following day. 3 - Dynamic Choice and Consumption

218