Informs Annual Meeting 2017

SC04

INFORMS Houston – 2017

SC03A Grand Ballroom A APS Student Paper Competition Sponsored: Applied Probability Sponsored Session Chair: Jose Blanchet, Columbia University, 500 West 120th Street, 340 Mudd Building, New York, NY, 10027, United States, jose.blanchet@gmail.com Co-Chair: John Hasenbein, University of Texas-Austin, 1 University Station Stop C2200, Department of Mechanical Engineering, Austin, TX, 78712-0292, United States, jhas@mail.utexas.edu 1 - Pricing and Optimization in Shared Vehicle Systems: An Approximation Framework Daniel Freund, Cornell University, 109 Lake St, Ithaca, NY, 14850, United States, df365@cornell.edu, Thodoris Lykouris Pricing in shared vehicle systems is challenging due to complex network externalities: altering prices in any location affects future supply throughout the system within short timescales. Such externalities are well captured by steady- state Markov chain models, which are widely used in this context. However, optimizing these models is computationally difficult since the resulting problems are high-dimensional and non-convex. We develop a framework for designing pricing policies in such systems, based on a novel convex relaxation which we term elevated flow relaxation, coupled with a new infinite projection and pullback technique for proving approximation bounds. Our approach provides the first efficient algorithms with finite-system guarantees for pricing as well as other demand-supply balancing controls used in shared vehicle systems. In the process, we recover recently discovered asymptotic optimality results via elementary arguments. 2 - On the Euler Discretization Error of Brownian Motion about Random Times Guido Lagos, University of Chile, Santiago, Chile, guido.lagos.barrios@gmail.com In this talk we study the simulation of barrier-hitting events and extreme events of Brownian motion, when using a discretization on an equidistant time mesh. Specifically, we study the times and position of the discretized Brownian motion in these events and compare it to the ones for the “real” Brownian motion. We establish new results on weak convergence of the (normalized) errors of time and position in all these cases, and give explicit analytic expressions for the limiting distributions. In doing this we derive new results on diffusion approximations of Gaussian random walks by Brownian motions. More importantly, our results give new insight on the connection between several works in the literature dating back to the 60’s where the constant $-\zeta(1/2)/\sqrt{2\pi}$ has appeared, where $\zeta$ is the Riemann zeta function. 3 - Multi-Agent Online Learning under Imperfect Information: Algorithms, Theory and Applications Zhengyuan Zhou, Stanford University, 160 Comstock Circle - Unit 106002, Stanford, CA, 94305, United States, zyzhou@stanford.edu We consider a model of multi-agent online learning under imperfect information, where the reward structures of agents are given by a general continuous game. After introducing a general equilibrium stability notion for continuous games, called variational stability, we examine the well-known online mirror descent (OMD) learning algorithm and show that the “last iterate” (that is, the actual sequence of actions) of OMD converges to variationally stable Nash equilibria provided that the feedback delays faced by the agents are synchronous and bounded. We then extend the result to almost sure convergence to variationally stable Nash equilibria under both unbiased noise and synchronous and bounded delays. Subsequently, to tackle fully decentralized, asynchronous environments with unbounded feedback delays, we propose a variant of OMD which we call delayed mirror descent (DMD), and which relies on the repeated leveraging of past information. With this modification, the algorithm converges to variationally stable Nash equilibria, with no feedback synchronicity assumptions, and even when the delays grow super-linearly relative to the game’s horizon. We then again extend it to the case where there are both delays and noise. 4 - Learning Preferences with Side-Information: Near Optimal Recovery of Tensors Andrew A. Li, Massachusetts Institute of Technology-ORC, 77 Massachusetts Avenue, Bldg. E40-149, Cambridge, MA, 02139, United States, aali@mit.edu Many recent problems in e-commerce can be cast as large-scale problems of tensor recovery. Thus motivated, we study the problem of recovering tensors from their noisy observations. We provide an efficient algorithm to recover structurally ‘simple’ tensors given noisy observations of their entries; our version of simplicity subsumes low rank tensors for various definitions of tensor rank. Our algorithm is practical for massive datasets and provides a significant performance improvement over incumbent approaches to Tensor recovery. Further, we show a near-optimal recovery guarantee. Experiments on music streaming data demonstrate the performance and scalability of our algorithm.

SC03C Grand Ballroom C

M&SOM Student Paper Competition II Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Tunay Tunca, University of Maryland, Silver Springs, MD, 20910, United States, ttunca@rhsmith.umd.edu 1 - Mallows-smoothed Distribution over Rankings Approach for Modeling Choice Antoine Desir, Columbia IEOR.Department, New York, NY, ad2918@columbia.edu, Vineet Goyal, Srikanth Jagabathula, Danny Segev Abstract not Available 2 - An Examination of Early Transfers to the ICU Based on a Physiologic Risk Score Wenqi Hu, Columbia Business School, New York, NY, Contact: wh2274@columbia.edu Abstract not Available 3 - Robust Dual Sourcing Inventory Management: Optimality of Capped Dual Index Policies Jiankun Sun, Northwestern University, Evanston, IL, Contact: jiankun-sun@kellogg.northwestern.edu Abstract not Available 320A Responsible Supply Chains and Sustainable Products Sponsored: Manufacturing & Service Oper Mgmt, Sustainable Operations Sponsored Session Chair: Vishal Agrawal, Georgetown University, Washington, DC, 20057, United States, va64@georgetown.edu 1 - Managing Social Responsibility in Multitier Supply Chains Robert Swinney, Associate Professor, Duke University, 100 Fuqua Dr, Box 90120, Durham, NC, 27708, United States, robert.swinney@duke.edu, Jing-Sheng Jeannette Song We study the management of social responsibility in a three-tier supply chain in which a Tier 2 supplier sells to a Tier 1 supplier, which in turn sells to a Tier 0 buyer. The Tier 2 supplier potentially violates social and environmental standards, resulting in harm to the Tier 0 and 1 firms. Each member of the supply chain can exert effort to improve the responsibility in Tier 2, and the efforts of Tiers 0 and 1 are substitutable with one another and complementary to the efforts of Tier 2. We determine when it is optimal for the Tier 0 buyer to delegate management of Tier 2 to the intermediate Tier 1 supplier, and when it is optimal for Tier 0 to assume direct control of Tier 2. 2 - Transparency as a Supplier Social and Environmental Responsibility Screening Mechanism Basak Kalkanci, Georgia Institute of Technology, 3151 Stillhouse Creek Dr SE, Apt 25517, Atlanta, GA, 30339, United States, basak.kalkanci@scheller.gatech.edu, Shi Chen, Erica Plambeck Some buying firms, facing scrutiny regarding negative social and environmental impacts in their suppliers’ operations, have recently made transparency commitments to disclose their audit reports and/or the identities of their suppliers. In this paper, we study the trade-offs faced by a buyer firm in deciding whether or not to make such transparency commitments. 3 - Improving Supplier Social Responsibility under Incomplete Visibility Leon Matias Valdes, MIT.Sloan School of Management, 77 Massachusetts Avenue, E62 - 459, Cambridge, MA, 02139, United States, lvaldes@mit.edu, Tim Kraft, Yanchong Zheng We study a manufacturer’s decisions when the social responsibility (SR) performance of her supplier cannot be perfectly observed. Only the supplier can directly impact SR; the manufacturer can: (i) invest in the supplier’s capabilities to improve SR; and (ii) disclose SR information (true or not) to consumers. In addition, a third party may disclose the true level of SR, which can lead to a penalty for the manufacturer. Our results provide insights into the impact of supply chain visibility on the supplier’s and manufacturer’s strategies. While greater visibility and third-party scrutiny typically increase SR in the supply chain, we identify situations where this is not the case. SC04

67

Made with FlippingBook flipbook maker