2015 Informs Annual Meeting

WE43

INFORMS Philadelphia – 2015

WE44 44-Room 103B, CC New Approaches in Dynamic Pricing and Revenue Management Sponsor: Revenue Management and Pricing Sponsored Session Chair: Yonatan Gur, Stanford GSB, 655 Knight Way, Stanford, CA, 94305, United States of America, ygur@stanford.edu 1 - Randomized Markdowns and Online Monitoring Ken Moon, PhD Candidate, Stanford GSB, 655 Knight Way, Stanford, CA, 94305, United States of America, kenmoon@stanford.edu, Kostas Bimpikis, Haim Mendelson Using data tracking customers of a North American retailer, we present empirical evidence that consumers are forward-looking and that monitoring products online associates with successfully obtaining discounts. Developing a structural model relating consumers’ dynamic behavior to their monitoring costs, we find substantial heterogeneity, with opportunity costs for an online visit ranging from $2 to $25 in inverse relation to price elasticities. We show implications for retail operations. 2 - Agent Behavior in the Sharing Economy: Evidence from Airbnb Antonio Moreno-Garcia, Northwestern University, 2001 Sheridan Rd, Evanston, Il, 60208, United States of America, a-morenogarcia@kellogg.northwestern.edu, Jun Li, Dennis Zhang Using data from Airbnb, we study the behavior of non-professional agents in two- sided platforms. 3 - Implications of Choice Paralysis on Operational Decision Making Rene Caldentey, NYU, 44 W 4th St, New York, NY, 10012, United States of America, rcaldent@stern.nyu.edu, Srikanth Jagabathula, Anisha Patel We empirically investigate the notion of choice paralysis (i.e., too many options can paralyze a consumer and make them more prone to not purchasing) and study its implications on assortment and inventory decisions. We propose a modification to the nested logit model to incorporate the choice paralysis effect. 4 - Doing While Learning and Adapting to a Changing Environment Yonatan Gur, Stanford GSB, 655 Knight Way, Stanford, CA, 94305, United States of America, ygur@stanford.edu, Omar Besbes, Assaf Zeevi Multi Armed Bandit (MAB) problems are building blocks of many RM&P problems. We study a MAB formulation that allows for a broad range of temporal uncertainties in the rewards. We characterize the complexity of this class of problems, mapping the ``budget” of allowable variation to the minimal achievable regret relative to a dynamic oracle. We study the price of universality: the additional complexity associated with not knowing variation budget, over the one embedded in a known budget. WE45 45-Room 103C, CC Reducing the Carbon Footprint Contributed Session Chair: Emre Berk, Bilkent University, Management Faculty, 06800 Bilkent, Ankara, Turkey, eberk@bilkent.edu.tr 1 - Real Options Portfolio Strategies for Cloud Infrastructure Expansion Yunpeng Pan, Assistant Professor, South Dakota State University, Mathematics&Statistics, Box 2220, Brookings, SD, 57007, United States of America, yunpeng.pan@sdstate.edu Cloud services are powered by capital-intensive, energy-hungry data centers. The temporal and spatial choices of data center deployment must be made judiciously to best satisfy customer needs while keeping economic and environmental costs in check. To this end, we propose a real options framework for evaluating the desirability of candidate sites under the complex dynamics of electricity rate, customer demand, etc.; we develop strategies for portfolio selection and option exercise.

3 - Institutional Logics Change and Firm Attention: Sustainability Logic in the Apparel Industry Yoojung Ahn, University of Massachusetts Amherst, MA, 121 Presidents Drive, Amherst, MA, 01003, United States of America, yoojung@som.umass.edu This paper explores how change in institutional logics impacts firm attention to this logic to create self-regulatory institutions. I examine the “sustainability logic” in the apparel industry to understand the different ways firms attend to this logic, and whether attention patterns contribute differently to participating in a self- regulatory institution. I deploy content analysis and event history analysis methods. 4 - Development of Predictive Model for Moviegoers using Multi Regression Analysis and Movie Scheduling Sung Wook Yun, Yonsei University, Sinchon-dong, Seodaemun-gu, Seoul, Korea, Republic of, giantguard@naver.com This paper is about a practical decision-making approach to a film screening in a multiplex movie theater. Our ultimate objective in this paper is to maximize the number of moviegoers by allocating movies to a limited number of screens that have different number of seats. We specifically devised a movie schedule model that determine which movies will be played on which screens with the consideration of an exchange screening and a double booking based on the Movie forecasting. Chair: Izak Duenyas, John Psarouthakis Professor, University of Michigan, Ross School of Business, Ann Arbor, MI, 48109, United States of America, duenyas@umich.edu 1 - Dual-metric Segmentation for Creating Airline Forecast Groups Wei Wang, Scientist, PROS, Inc., 3100 Main Street, Suite #900, Houston, Tx, 77002, United States of America, weiwang@pros.com For airlines, forecast groups created based on various flight attributes can improve forecast accuracy and provide sponsorship (especially to new markets), however frequently the segmentation uses load factor (LF) as the only metric. We present a two-metric and two-step approach where the segmentation is guided by both revenue and LF. 2 - Analysis of Self-adjusting Controls for Dynamic Pricing with Unknown Demand Parameters George Chen, Stephen M. Ross School of Business, University of Michigan, 701 Tappan Ave, Ann Arbor, MI, United States of America, georgeqc@umich.edu, Izak Duenyas, Stefanus Jasin We study the network-RM pricing problem with unknown demand function parameter. We develop a joint learning and dynamic pricing heuristic that combines MLE and self-adjusting price control and show that the best attainable revenue loss rate in the general setting can be achieved without re-optimization. A much sharper rate can also be achieved when demand are well-separated using the proposed self-adjusting heuristic. 3 - Data-driven Algorithms for Nonparametric Multi-product Inventory Systems Weidong Chen, University of Michigan, Industrial and Operations Engineering, Ann Arbor, MI, 48109, United States of America, aschenwd@umich.edu, Cong Shi, Izak Duenyas We propose a data-driven algorithm for the management of stochastic multi- product inventory systems with limited storage as well as production cost uncertainty. The demand distribution is not known a priori and the manager only has access to past sales data. We measure performance of our proposed policy through regret and characterize the rate of convergence guarantee of our algorithm. 4 - Dynamic Pricing and Inventory Management under Network Externalities Renyu Zhang, Doctoral Student, Olin Business School, Washington University in St. Louis, Campus Box 1133, 1 Brookings Drive, St. Louis, MO, 63130, United States of America, renyu.zhang@wustl.edu, Nan Yang We study a periodic-review joint pricing and inventory management model with network externalities. The product is a network product so that the customers’ willingness-to-pay and, thus, the potential demand of the product are increasing in the network size. We characterize the optimal policy and analyze the impact of network externalities upon the optimal price and inventory decisions. We also propose effective strategies to exploit network externalities. WE43 43-Room 103A, CC Pricing and Inventory Control Sponsor: Revenue Management and Pricing Sponsored Session

492