2015 Informs Annual Meeting

MD45

INFORMS Philadelphia – 2015

MD46 46-Room 104A, CC Equilibrium Models and Pricing of Queues Sponsor: Manufacturing & Service Oper Mgmt/Service Operations Sponsored Session Chair: Philipp Afeche, Associate Professor, University of Toronto, 105 St. George Street, Toronto, ON, M5S3E6, Canada, afeche@rotman.utoronto.ca 1 - An Equilibrium Analysis of a Multiclass Queue with Endogenous Abandonments Xiaoshan Peng, The University of Chicago, Booth School of Business, 5807 S Woodlawn Ave, Chicago, Il, 60637, United States of America, x-peng@chicagobooth.edu, Baris Ata, Peter Glynn This paper studies a multiclass queueing system with endogenous abandonments where the congestion affects customers’ abandonment behavior and vice versa. Our model captures this interaction by developing two closely related models: an abandonment model and a queueing model. Combining the results for the two models, we show that there exists a unique equilibrium in which the customers’ abandonment time and the virtual waiting time for the various classes are consistent in the two models. 2 - Observational Learning and Abandonment John Yao, Columbia University, 3022 Broadway, New York, NY, 10027, United States of America, JYao14@gsb.columbia.edu, Costis Maglaras, Assaf Zeevi Demand models used in service operations often assume that users have accurate knowledge of the service system parameters needed to make decisions, such as whether to join a queue. What if instead, users must form estimates of these system parameters through their own observations or experiences in the system? I show the effect of observational learning on user behavior and equilibrium system performance in the context of abandonment in a queueing model. 3 - Strategic Open Routing in Queueing Networks Andrew Frazelle, Fuqua School of Business, Duke University, 100 Fuqua Drive, Durham, NC, 27708, United States of America, andrew.frazelle@duke.edu, Alessandro Arlotto, Yehua Wei Motivated by self-interested routing, we propose a two-station queueing network with open routing in which agents require service at both stations and seek to minimize their individual total system times. Agents have no inherent preference over the sequence of stations that they visit apart from the impact of this sequence on their system times. We evaluate system performance and determine Nash or subgame perfect Nash equilibrium customer routing behavior in three different overloaded settings. 4 - Pricing and Prioritizing Time-sensitive Customers with Heterogeneous Demand Rates Providers often face time-sensitive customers that differ in their demand rates. Examples include amusement parks, museums, and ski resorts. However, the pricing literature for queues typically assumes unit demand for all customers. We study a revenue-maximizing provider that designs a price/lead-time menu for customers with heterogeneous demand rates and private information on their preferences. We show under what conditions it is optimal to prioritize customers based on their demand rates and transaction values, even if all are equally time- sensitive. Ricky Roet-Green, University of Toronto, 37 Zola Gate, Thornhill, L4J9A7, ON, Canada, rgricky@gmail.com, Opher Baron, Philipp Afeche, Joseph Milner

2 - Pay-as-You-go Business Models for Energy Technology Innovations in Developing Economies Jose Guajardo, University of California Berkeley, 545 Student

Services Bldg #1900, Berkeley, CA, 94720-1900, United States of America, jguajardo@berkeley.edu

Pay-As-You-Go business models have become widespread for the diffusion of energy technology innovations in developing economies, yet not much is known about this recent phenomenon. In this research, we analyze central aspects of consumer behavior and contract design in these novel markets. 3 - Selling Freemium Products to Loss Averse Consumers Sami Najafi-Asadolahi, Santa Clara University, 500 El Camino Real, Santa Clara, CA, United States of America, snajafi@scu.edu, Nishant Mishra, Andy Tsay We consider a firm selling two versions of a single product, a freemium for free and a premium at a regular price, to consumers who are loss-averse. Each consumer first uses the freemium, and after using it, decides whether to buy the premium. We find that when consumers become slightly dissatisfied from the freemuim’s valuation they, counter-intuitively, become more willing to purchase the premium, thereby increasing the firm’s revenue. 4 - Product Recommendations via Geometric-based Adaptive Choice Conjoint Analysis Denis Saure, University of Chile, Republica 701, Santiago, Chile dsaure@gmail.com, Juan Pablo Vielma Aiming to obtain individualized estimates of consumer preferences in the context of product recommendations, we study the construction of adaptive conjoint choice designs under a Bayesian framework. By adopting a geometric interpretation of the problem, we construct near optimal designs when the number of questions is small, and also give a precise interpretation of efficiency criteria and design methods used in extant research, which we show result in suboptimal designs. Chair: John Birge, Professor, University of Chicago Booth School of Business, 5807 S Woodlawn Ave, Chicago, IL, 60637, United States of America, john.birge@chicagobooth.edu 1 - Learning to Compete Against Dynamic Pricing Strategies Matthew Stern, University of Chicago - Booth School of Business, 5807 S Woodlawn Ave, Chicago, IL, 60637, United States of America, stern@chicagobooth.edu, John Birge We examine the impact of competition on the design and implementation of dynamic pricing strategies. Observing posted prices in each period, as well as their own private demand realizations, firms compete for revenues while learning the parameters of their underlying demand curves. We show that when firms remain willfully ignorant of their business environment, they can sustain collusive prices. We conclude that incomplete learning is a desired outcome of competing dynamic pricing strategies. 2 - Dynamic Pricing and Learning in Spread Betting Adam Schultz, PhD Student, University of Chicago Booth School of Business, 5807 S Woodlawn Ave., Chicago, IL, 60637, United States of America, adam.schultz@chicagobooth.edu, John Birge, Bora Keskin We develop a model in which a sportsbook dynamically prices the point spread for a sporting event. In our model, bettors maximize their expected profits by timing their bets, while the sportsbook follows an easily implementable policy to update the point spread. To analyze the decisions of the betting market, we introduce a mean-field approximation. Using data from online sportsbooks, we reveal insights about the betting market. 3 - Rental System Revenue Management Problem with Totally Unimodular Constraints Ali Cem Randa, University of Chicago Booth School of Business, We analyze the example of a renter which has finite number of identical units that can be loaned for durations of days. The renter has to determine its booking limits for a planning horizon of finite duration which is considerably far in the future. We assume that all permutations of consecutive days in the planning horizon define a different product. The capacity constraints formed by these products are totally unimodular. We solve a multi-stage stochastic program exploiting this structure. 5807 Woodlawn Ave., Chicago, IL, 60637, United States of America, randa@chicagobooth.edu, John Birge, Baris Ata MD45 45-Room 103C, CC Revenue Management and Learning II Sponsor: Revenue Management and Pricing Sponsored Session

246