2015 Informs Annual Meeting

SA53

INFORMS Philadelphia – 2015

SA54 54-Room 108A, CC Data-Driven Stochastic Programming using Phi-Divergences

2 - Impact of Top Service Designers on Experience Service Performance

Gregory Heim, Associate Professor, Mays Business School at Texas A&M University, Wehner Hall 320, College Station, TX, 77843- 4217, United States of America, GHeim@mays.tamu.edu Service firms today leverage many designers known for design excellence. Despite expanding use of top designers, no research examines practical implications of top designers as compared to mainstream designers. In considering whether to use top designers, service managers must consider benefits from investment in high quality designs. Using data on golf courses and golf course designers, we study how top designers influence performance of Texas golf courses. 3 - Applying Normalized Systems Theory to Service System Design Ralph Badinelli, Professor, Virginia Tech, Dept. of Busines Information Technology, Virginia Tech 0235, Blacksburg, VA, 24061, United States of America, ralphb@vt.edu This paper applies Normalized Systems Theory (NST) as a new framework for the modular design of service systems. We synthesize a prescription for service system design from principles of NST and principles of the Viable Systems Approach (VSA). A mathematical model of the effects of system coupling on evolvability and entropy is derived. Evolvability and entropy are related to autopoiesis, homeostasis and viability. An example of an NST-inspired service system design is provided. SA53 53-Room 107B, CC Experiments in Supply Chains Sponsor: Behavioral Operations Management Sponsored Session Chair: Andrew Davis, Cornell University, 401J Sage Hall, Ithaca, NY, John Aloysius, University of Arkansas, WCOB 475d University of Arkansas, Fayetteville, AR, 72701, United States of America, JAloysius@walton.uark.edu, Siqi Ma Research on inventory decisions under uncertainty has primarily focused on aleatory uncertainty due to stochastic variability. Epistemic uncertainty due to a lack of confidence in knowledge however routinely features in such decisions. Our experiment orthogonally manipulates the two forms of uncertainty in a newsvendor task. We partial out biases due to each form of uncertainty and also report on learning as well as interaction effects. 2 - Bargaining and the Allocation of Risk in Supply Chains: An Experimental Study Kyle Hyndman, University of Texas at Dallas, 800 W Campbell Rd (SM 31), Richardson, TX, 75080, United States of America, KyleB.Hyndman@utdallas.edu, Andrew Davis We study the impact of bargaining and inventory risk location in a supply chain. We conduct a human-subjects experiment where a retailer and supplier engage in free-form bargaining over wholesale price, order quantity and inventory location. We show: (1) our bargaining environment leads to higher efficiency than in past studies; (2) the party incurring the inventory risk always earns a substantially lower profit than the other party; (3) more restrictive environments lead to lower efficiency. 3 - Sharing in the Benefits of Learning-by-doing: A Laboratory Study of Procurement Auction Mechanisms Blair Flicker, PhD Candidate, The University of Texas at Dallas, 800 W Campbell Road, Mailstop AD23, Richardson, TX, 75208, United States of America, bflicker@utdallas.edu, Wedad Elmaghraby, Elena Katok Buyers looking to share in suppliers’ learning-by-doing savings can either (i) draft long-term contracts with payment reduction schedules or (ii) encourage competition via sequential short-term contracts. Theoretically, there is minimal cost difference between the two approaches, but behavioral findings suggest that humans do respond differentially to the two mechanisms. 4 - Supply Base Diversification in the Presence of High Impact, Low Probability Supply Disruptions Doug Thomas, Penn State, 463 Business Building, University Park, PA, 16802, United States of America, dthomas@psu.edu, Mirko Kremer, Kyle Goldschmidt, Chris Craighead We investigate sourcing decisions in the presence of high impact, low probability supply disruptions. We develop a model that captures a key sourcing tradeoff: A consolidated (diversified) supply base reduces (increases) transaction costs but increases (reduces) the exposure to disruptions. We predict, and using a laboratory experiment, find evidence for, an oscillating pattern - decision makers diversify immediately after a severe disruption and consolidate during stretches without disruptions. 14850, United States of America, adavis@cornell.edu 1 - Inventory Decisions under Epistemic and Aleatory Demand Uncertainty

Cluster: Tutorials Invited Session

Chair: Guzin Bayraksan, Associate Professor, The Ohio State University, Integrated Systems Engineering, Columbus, OH, 43209, United States of America, bayraksan.1@osu.edu 1 - Data-Driven Stochastic Programming using Phi-Divergences Guzin Bayraksan, Associate Professor, The Ohio State University, Integrated Systems Engineering, Columbus, OH, 43209, United States of America, bayraksan.1@osu.edu, David K. Love Phi-divergences provide a measure of distance between two probability distributions. They can be used in data-driven stochastic optimization to create an ambiguity set of distributions centered around a nominal distribution and hedge against distributional uncertainty. In this tutorial, we present two-stage models with distributional uncertainty using phi-divergences and tie them to risk-averse optimization. We examine the value of collecting additional data and discuss convergence properties. SA55 55-Room 108B, CC Airline Economics: Competition and Collaboration Sponsor: Aviation Applications Sponsored Session Chair: Yi Liu, UC Berkeley, 107 McLaughlin Hall, Berkeley, Ca, 94720, United States of America, liuyi.feier@gmail.com 1 - Airline Competition and Market Frequency: A Comparison of the S-curve and Schedule Delay Models Yi Liu, UC Berkeley, 107 McLaughlin Hall, Berkeley, CA, 94720, United States of America, liuyi.feier@gmail.com, Mark Hansen We compare two models for an airline market served by identical carriers who compete on fare and frequency: s-curve model and schedule delay model. They only differ structurally with respect to how they handle frequency competition. The demand side of our model is an approximation of a nested logit model which yields endogenous travel demand by including not travelling in the choice set. The results of our comparison support the schedule delay model over the s-curve model. 2 - Game-theoretic and Empirical Models for Airline Capacity and Fare Decisions under Competition Vikrant Vaze, Assistant Professor, Dartmouth, 14 Engineering Drive, Hanover, NH, 03755, United States of America, Vikrant.S.Vaze@dartmouth.edu, Reed Harder Airline capacity and fare decisions under competition affect passenger choice and the overall efficiency of airline networks. We develop Nash equilibrium models as well as empirical models to characterize these interactions. We analytically prove desirable properties of the resulting games and describe numerical experiments that extend our theoretical results to more complicated competition settings. Our results, validated against real-world data, provide new insights into airline competition. 3 - An Equitable and Collaborative Mechanism for Scheduling Interventions at Congested Airports Alexandre Jacquillat, PhD Candidate, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Building E40-240, Flight scheduling interventions can mitigate airport congestion by controlling the imbalances between peak-hour demand and capacity. We design, optimize and assess non-monetary congestion-mitigating scheduling mechanisms that ensure inter-airline equity and enable airline collaboration. Theoretical and computational results suggest that large equity gains can be achieved at small efficiency losses and that accounting for airline preferences can improve the outcome of scheduling interventions. Cambridge, MA, 02116, United States of America, alexandre.jacquillat@gmail.com, Vikrant Vaze

56