Informs Annual Meeting 2017

TA04

INFORMS Houston – 2017

3 - Strategic Inventories when Competing Manufacturers Sell through a Common Retailer Abhishek Roy, University of Texas-Austin, 2110 Speedway Stop B6500, Austin, TX, 78712-1277, United States, abhishek.roy@utexas.edu, Stephen M. Gilbert, Guoming Lai Although the effects of strategic inventory in dynamic contracts on supply chain agents are well known, most existing research on strategic inventory focuses on bilateral monopolies. When the partially substitutable products of different manufacturers are sold through a common retailer, we find that those effects are altered in significant ways. Among other things, the retailer’s and the manufacturers’ preferences between commitment and dynamic contracts can be reversed. 4 - Distribution Channel Entry Strategies with Intra-and Inter-channel Demand Learning Shuya Yin, University of California-Irvine, Paul Merage School of When a rm introduces a new product to market, making it available in multiple channels can achieve higher market coverage and learn market uncertainty better so as to improve subsequent decisions, while selling it in a single channel exclusively can save slotting fees and avoid channel cannibalization. This paper considers a model where a rm needs to determine its distribution channel entry strategy on whether to sell its product in one or two channels over a two-period time. 5 - Incentives for Information Transparency under Vertical Information Asymmetry Sammi Tang, Associate Professor, University of Miami, Miami, FL, 33146, United States, ytang@miami.edu, He Huang, Hongyan Xu This paper studies the incentives for horizontal information transparency between competing downstream firms when upstream suppliers possess private information. Information is modeled along two dimensions: vertical information on suppliers’ reliability level and horizontal information on suppliers’ production cost. Our findings provide novel insights for managers to handle two-dimensional information asymmetry. 320A Sustainable Operations Sponsored: Manufacturing & Service Oper Mgmt, Sustainable Operations Sponsored Session Chair: Greys Sosic, University of Southern California, Los Angeles, CA, 90089-0809, United States, sosic@marshall.usc.edu Co-Chair: Hailong Cui, University of Southern California, Los Angeles, CA, 90089-0809, United States, Hailong.Cui.2019@marshall.usc.edu 1 - Lemons, Trade-ins and Remanufacturing Ximin Huang, Assistant Professor, University of Minnesota, 321 19th Ave S, Ste 3-150, Minneapolis, MN, 55455, United States, huangx@umn.edu, Atalay Atasu, Beril L.Toktay Trade-in programs have been shown to partially mitigate the lemons problem in secondary markets. We show when and how remanufacturing traded-in products can further improve the efficiency in secondary markets. 2 - Complementary Product Design and Process Technology Choices in Recycling Luyi Gui, The Paul Merage School of Business, University of California-Irvine, Irvine, CA, 92697-3125, United States, luyig@uci.edu, Morvarid Rahmani, Atalay Atasu Motivated by the Japanese implementation of the Home Appliance Recycling Law, we study the complementarity between product design and technology choices for recycling. We evaluate the environmental benefit of recycling capacity sharing between producers and of producers’ investment in recycling facilities. TA04 Business, Irvine, CA, 92697-3125, United States, shuya.yin@uci.edu, Jing Hu, Christopher S.Tang

grids. Due to increasing uncertainty, traditional reliability constraints must be replaced by probabilistic guarantees. We investigate the probability of line current overloads assuming stochastic power injections, using techniques from large deviations theory and concentration inequalities, both in a static and in a dynamic scenario. We derive analytic expressions for the set of admissible power injections profiles such that line overloads are sufficiently rare events, and we show how enforcing constraints on line temperature, rather than on current, can lead to capacity gains. 2 - Learning to Price in Demand Response Programs Kia Khezeli, Cornell University, Ithaca, NY, 14853, United States, kk839@cornell.edu, Eilyan Bitar Electric power utilities offer demand response (DR) programs where consumers are paid to curtail their electricity consumption when electricity is scarce and expensive. The most basic challenge that a utility faces in implementing DR programs is the prediction of how customers will adjust their demand in response to different prices. In this talk, we’ll discuss the role of adaptive control in guiding the design of DR pricing polices that effectively balance the tradeoff between the utility’s need to learn the unknown customer response model and the need to minimize its expenditure over time. 3 - Estimation of Stochastic Process Parameters Implied by Financial Transmission Rights Boris Defourny, Lehigh University, 200 W Packer Ave, Bethlehem, PA, 18015, United States, defourny@lehigh.edu This work hypothesizes that Financial Transmission Right (FTR) prices in electricity markets express the assessment by market participants of the grid congestion risk, as determined by parameters of stochastic processes such as load levels, generation and transmission grid state driving the underlying day-ahead hourly price differences. We propose a methodology to infer parameters consistent with FTR contract prices, based on a relation between state occupation measures and the contract value. More broadly, this work is concerned with the inverse mapping from occupation measures to stochastic processes, and the stability of the estimates. 4 - Random Graph Models for Power System Controllability Linqi Guo, California Institute of Technology, Pasadena, CA, United States, lguo@caltech.edu, Steven H.Low We give a full characterization in terms of spectral graph theory for the controllability of the swing and power flow dynamics when we can only inject control signals to a subset of the buses. In particular, we show the controllability of the system depends on two orthogonal conditions: 1) intrinsic structure of the system graph 2) algebraic coverage of buses with controllable loads. Both conditions can be verified using the eigenvalue and eigenvectors of the graph Laplacian matrix. We then employ random graph models and uncertainty principle to explain why in practice we only need sporadic controllers for full controllability. TA03C Grand Ballroom C Economic Models on Markets, Firms and Information Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Shuya Yin, University of California-Irvine, shuya.yin@uci.edu 1 - When Does a Supply Chain Member Benefit from Vendor Managed Inventory? Ruixia Shi, University of San Diego, San Diego, CA, United States, rshi@sandiego.edu, Jun Ru, Jun Zhang This paper studies the effects of VMI on a supply chain consisting of one manufacturer and one retailer. We show that whether the two supply chain members benefit from VMI depends on how the holding or shortage cost increment from the manufacturer to retailer compares with two corresponding critical values. We then develop comparative statics results on how these critical values change with respect to different parameters. Interestingly, the retailer is more likely to benefit from the adoption of VMI when its inventory holding cost is low, and the manufacturer is more likely to benefit from VMI adoption when its inventory holding cost is high, contradicting what our intuitions would suggest. 2 - Inventory Allocation in the Presence of Store Switching Caused by Stock Outs Elcin Ergin, McGill, 3465 Hutchison Street, Apt 905, Montreal, QC, H2X 2G3, Canada, elcin.ergin@mail.mcgill.ca, Mehmet Gumus In this study, we focus on inventory allocation of fast fashion products between the stores of the firm when there is an evidence of store switching behaviour of customers in the case of stock outs. Examining an extensive dataset taken from a fast fashion retailer company which has more than 300 stores showed that some of the customers visit other stores to find the product they look for. We present an inventory allocation model considering this behavior. We analyze how this behavior would affect the total sales and accordingly how a firm should shape the initial allocation decisions under this model.

269