Speaker: Stephen C. Graves , Abraham J. Siegel Professor of Management Science at MIT
Title: Optimal Planning of a Reverse-Logistics Supply Chain for a Consumer Electronics Retailer
Abstract:We describe, model, and optimize the reverse logistics of our industrial partner, a Fortune 100 company that sells consumer electronics. For this company, as for many other retail businesses, the costs associated with the management of warranty claims and regret returns can be significant. This is further complicated by the short life cycle of the products, usually less than 1 year. A key planning challenge is to determine how much inventory is needed to support warranty exchanges, accounting for the uncertainty in the return rates over the product life cycle, the time to refurbish failed units, and the rate at which inventory depreciates in value. Within this setting, our main contributions are: (i) A forecasting strategy that focuses on estimating time-dependent failure distributions of products, allowing for priors constructed using historical data; (ii) A discrete-time stochastic model for inventory management that captures the dynamic behavior of this system; (iii) An algorithm for assigning refurbished items in inventory to customers that have made warranty claims. All the results will be illustrated through numerical experiments using data provided by our partner.
Speaker: Steve Gilbert, McCombs School of Business, University of Texas at Austin
Title: Product Line Extensions and Technology Licensing with a Strategic Supplier
Abstract:In many industries, original equipment manufacturers (OEMs) must obtain critical compo- nents from a few powerful suppliers. For example, OEMs that produce information technology hardware typically interact with highly concentrated supply industries that are dominated by a few key participants, e.g. Microsoft, Intel, etc. To the extent that the OEMs are also con- centrated, e.g. Dell, Hewlett-Packard, etc., the interactions between the suppliers of critical components and the OEMs are strategic, and have implications for how an incumbent OEM chooses its product line and interacts with potential rivals. We demonstrate that, by adding a low-end product line extension, an OEM can induce a common supplier to o_er more favorable pricing. Moreover, depending upon the cost structure and relative performance of the product line extension, the OEM may bene_t even more from the low-end line extension if it is produced by a rival instead of by itself, even if it cannot obtain any licensing income from it. Among other things, we show that this can result in a decentralized OEM accommodating competition from rivals producing product line extensions that would not be developed in a vertically integrated supply chain.
Key words: game theory; marketing strategy; competition; vertical di_erentiation; strategic e_ect
Speaker:Justin Jia, Krannert School of Management, Purdue University
Title: Mitigating American Drug Shortages: Contract Design and Operational Strategies
Abstract:Drug shortages have recently been a critical problem faced by the pharmaceutical industry. In this study, we investigate the reasons for the shortages, derive the optimal inventory policies for drug manufacturers, and design efficient contracts to mitigate the shortages. Our models capture key characteristics of drug shortages, including frequent manufacturing disruptions, uncertain recovery duration from disruption, limited buyers' patience, low redundancy in capacity, and weak failure-to-supply clauses. Our analytical and numerical analyses provide insights and managerial recommendations on drug shortage mitigation.
*This is a joint work with Hui Zhao, The Smeal College of Business, The Pennsylvania State University, State College, PA 16802
Speaker:John Buzacott, Schulich School of Business, York University ,Toronto, Canada
About the Speaker:John Buzacott was born in Sydney, Australia. He graduated from the University of Sydney with degrees in Physics and in Electrical Engineering. He then worked and studied in the UK, obtaining an M.Sc. in Operational Research and a PhD in Engineering Production from the University of Birmingham. He moved to Canada in 1967 and taught at the University of Toronto, the University of Waterloo and York University. He has been President of the Canadian Operational Research Society, President of the Production and Operations Management Society (POMS), and Chair, ORSA Special Interest Group on Manufacturing and Operations Management (the predecessor society of the INFORMS MSOM society). He is a Fellow of INFORMS, a Fellow of POMS, and a Fellow of MSOM. He is the co-author of four books and the author of over 100 publications in scientific journals. In 2001 he was awarded the degree of Doctor Honoris Causa by the Technical University of Eindhoven in the Netherlands.
Title: The Evolution of OR: From Operational Research to OM and Policy Analysis
Abstract:This talk is a personal perspective on how OR has evolved over the speaker’s 50 year career. I talk both about OR’s achievements and about its unmet promises. Particular emphasis will be given to the evolution of both the research process and the way in which OR is practiced. The talk will conclude with some thoughts on how OR might develop in the future.
Speaker:Haresh Gurnani, Professor of Management and Leslie O Barnes Scholar at the University of Miami
Title: Price Competition with Optimal Product Demonstrations
Abstract:We develop a game theoretic model of price competition in which an innovating firm can offer product demonstrations. Placing minimal restriction on the firm's ability to design demonstrations, we show that the equilibrium demonstration resolves some but not all customer valuation uncertainty and allows the innovating firm to attract customers while maintaining a high price. Consumer surplus may be lower with endogenous demonstrations than without demonstrations. Regulation requiring firms to provide fully-informative demonstrations (e.g., generous return policies or inspection periods) can further reduce consumer surplus. The ability to design demonstrations also creates incentives for innovating firms to limit the market appeal of their products, suggesting another mechanism through which product demonstrations can reduce market efficiency. The results have implications for firm management and pricing strategies and for consumer protection. Keywords: price competition, Bayesian persuasion, product demonstrations, trial periods, return policies, test drives
Speaker: Gemma Berenguer Falguera, Krannert School of Management, Purdue University
Title: A Capacitated Facility Location Model with Bidirectional Flows
Abstract:Supply chains with returned products are receiving increasing attention in the operations management community. The present paper studies a capacitated facility location model with bidirectional flows and marginal value of time for returned products. The distribution system consists of a single supplier that provides one new product to a set of distribution centers (DCs), which then ship to the final retailers. While at the retailers’ site, products can be shipped back to the supplier for reprocessing. Each DC is capacitated and handles stocks of new and/or returned product. The model is a nonlinear mixed-integer program that optimizes DC location and allocation between retailers and DCs. We show that it can be converted to a conic quadratic program, which can be efficiently solved. Some valid inequalities are added to the program to improve computational efficiency. We conclude by reporting numerical experiments that reveal some interesting properties of the model.
Speaker: J. George Shanthikumar, Krannert School of Management, Purdue University
Title: Structural Properties of Stochastic Inventory Systems
Abstract:We will review the single-period price-setting newsvendor models with reliable (see Petruzzi and Dada, 1999) and unreliable (see Dada et. al., 2007) suppliers. Following this, we will review the multi-period inventory control problems with pricing and reliable (see Federgruen and Heching, 1999) and unreliable (see Feng 2010 and Feng & Shi, 2012) suppliers. Our focus will be on the properties of the demand and supply functions that are needed to have specific ordering structures such as base stock or threshold policies.
Speaker: Thanh Nguyen, Krannert School of Management, Purdue University
Title: Local Bargaining and Endogenous Fluctuations
Abstract:We study how local bargaining in a networked market can cause endogenous fluctuations by a new approach that incorporates non-cooperative bargaining into a large networked economy. In particular, we consider a networked bargaining game that captures trade with intermediaries and define its replications. We examine the agents’ behavior in the limit as the population size goes to infinity: a limit stationary equilibrium exists if there is a converging sequence of semi-stationary equilibria in the finite replications. The existence of a limit stationary equilibrium captures the hypothesis that when the market gets large, the agents will behave myopically and the market will be stable. However, we prove that limit stationary equilibria need not exist even when market fundamentals is deterministic, agents are patient and share common beliefs. This shows that in our setting the underlying network is the main friction that hinders stationary markets.
Speaker: Yaroslav Rosokha, Krannert School of Management, Purdue University
Title: Learning under Compound Risk vs. Learning under Ambiguity - An Experiment
Abstract: We design and conduct an economic experiment to investigate the learning process of the agents under compound risk and under ambiguity. We gather data for subjects choosing between lotteries involving risky and ambiguous urns. Decisions are made in conjunction with a sequence of random draws with replacement, allowing us to estimate the beliefs of the agents at different moments in time. For each of the urn types we estimate the initial prior and a model of Bayesian updating allowing for base rate fallacies. Our findings suggest an important difference in updating behavior between risky and ambiguous environments. Specifically, after controlling for the initial prior, when updating under ambiguity, subjects significantly underweight the new signal, while when updating under compound risk subjects are essentially Bayesian.
Speaker: Masha Shunko, Krannert School of Management, Purdue University
Title: Pareto Improving Coordination Policies in Queueing Networks: Application to Flow Control in Emergency Medical Services
Abstract: One of the well known methods to improve performance in a queueing network is implementing some coordination policy that balances the load between servers. However, in decentralized queueing networks where each service agent can decide whether to participate in the system or not, coordination policy has to not only benet the system, but should also be benecial to all agents. In particular, agents are willing to participate in a coordination policy only if the performance of their queueing system is not hindered and if their revenues are not decreased. As a motivating example, we use the emergency medicine setting, in which emergency departments (EDs) act as independent agents and overcrowding in the EDs has direct impact on the quality of service. In such setting, agents are interested in seeing improvements in performance measures that address the risk of having an overcrowded ED in addition to the improvements in expected wait time in queue (which is a widely studied and applied metric). We focus on reducing the expected size of the queue, the variance of the queue size, the probability of having a long queue, and the expected queue size in the overcrowded state; and propose classes of coordination policies that provide improvement on all of these measures for all agents. In addition, agents who receive revenue based on the processed load, are interested in preserving the expected load. For example, in healthcare systems, ambulance diversion policies that attempt to balance the patient load between neighboring EDs in an area have not been very successful partially because the EDs are not always willing to divert ambulances due to potential revenue losses. Hence, our proposed classes of coordination policies guarantee that the expected arrival rate and hence, the expected revenue, is preserved for each agent in the network.