Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MC26

4 - Algorithmic Trading with Partial Information: A Mean Field Game Approach Philippe Casgrain, Sevastian Jaimungal, University of Toronto, Toronto, ON, Canada. Financial markets are often driven by latent factors which traders cannot observe. Here, we address an algorithmic trading problem with collections of heterogeneous agents who aim to perform statistical arbitrage, where all agents filter the latent states of the world, and their trading actions have permanent and temporary price impact. This leads to a large stochastic game with heterogeneous agents. We solve the stochastic game by investigating its mean-field game (MFG) limit, with sub-populations of heterogeneous agents, and, using a convex analysis approach, we show that the solution is characterized by a vector-valued forward- backward stochastic di erential equation (FBSDE). We demonstrate that the FBSDE admits a unique solution, obtain it in closed-form, and characterize the optimal behaviour of the agents in the MFG equilibrium. Moreover, we prove the MFG equilibrium provides an -Nash equilibrium for the finite player game. We conclude by illustrating the behaviour of agents using the optimal MFG strategy through simulated examples. 5 - Dynamic Mean-Risk Asset Allocation and Myopic Strategies: A Universal Portfolio Rule Zhaoli Jiang, The Chinese University of Hong Kong, Hong Kong, China, Xue Dong He In a market that consists of multiple stocks and one risk-free asset whose expected return rates and volatility are deterministic, we study a continuous-time mean-risk portfolio selection problem in which an agent is subject to a constraint that the expectation of her terminal wealth must exceed a target and minimizes the risk of her investment, which can be the variance or tail risk of her terminal wealth. Setting the target to be proportion to the agent’s current wealth, we derive the equilibrium policy in closed form, and this policy is myopic and does not depend on the risk measure used by the agent nor on the agent’s evaluation period. For another two targets, one that is the risk-free payo of the agent’s current wealth plus a premium and the other that is a weighted average of the risk-free payo of the agent’s current wealth and a pre-determined target, we also derive the equilibrium policy in closed form when the agent measures risk by the variance of her terminal wealth. n MC24 North Bldg 131B E-Business Sponsored: EBusiness Sponsored Session Chair: Varun Gupta, Penn State Erie, The Behrend College, Erie, PA, 16563, United States 1 - Role of Trust in Peer-to-peer Sharing Economy Jagan Jacob, University of Rochester, Simon Business School, 4-349 Carol Simon Hall, Rochester, NY, 14627, United States In online-based sharing-economy platforms such as Airbnb and Uber, trust between users is critical. If there are some “bad users who are discourteous with improper behavior, some risk-averse users may leave the platform. Stricter screening and entry requirements can reduce the probability of “bad users entering the platform. But, it also negatively affects the entry decisions of “good users, because of high entry-costs. How easy should it be for new users to sign-up to use the platform? What degree of background checking and screening procedure should the platform employ? I try to answer these questions using population dynamics models used in Ecology. 2 - Multiproduct Dynamic Upgrades Xiao Zhang, The University of Texas at Dallas, 800 W. Campbell Rd., SM 30, Richardson, TX, 75080, United States, Metin Cakanyildirim, Ozalp Ozer Upgrades in travel industry are often static and offered either at the booking time or at the check-in time. In this paper, we study dynamically-offered upgrades by a multi-product firm via notifications (e.g., emails) between the booking and the check-in times. We investigate a general multi-level upgrade policy in which a customer may be upgraded to any better products and a restrictive single-level upgrade policy which is less computationally intensive. Both policies have clean structures and are easy to implement. We also identify intuitive monotonicity properties for the optimal single-level upgrade policy. 3 - Demand Throttling for Bandwidth Preservation Varun Gupta, Penn State-Erie, The Behrend College, 5101 Jordan Rd, Burke 281, Erie, PA, 16563, United States, Sandun Perera Demand throttling is a commonly observed phenomenon for online service providers, who maintain their demand below a certain threshold to avoid any potential server crashes and to preserve their bandwidth. Using a generalized demand model based on Brownian motion, we analytically show that the optimal throttling policy resembles the (s,S) policy which is a well-known and commonly used inventory policy. Next, we provide numerical analysis and comparative statics of the analytical results. Finally, we discuss the managerial insights for bandwidth throttling policies in practice for online service providers.

4 - Online Order Fulfillment with Central and Local Warehouses when Facing Strategic Customers Chao Liang, Cheung Kong Graduate School of Business, Main Campus, Oriental Plaza 2/F,, Tower E2, 1 East Chang An Ave.,, Beijing, 100738, China, Yuxin Chen The e-tailer faces the problem of where to and how to allocate the inventory. If it keeps the inventory in suburban area (and thus far away from the customers), then it can enjoy a low warehouse renting cost but the shipping cost would be high. If it puts the inventory close to the customers, then it can save the shipping cost but the warehouse renting cost would be high. Another option is to keep inventory in both local and remote places. We study the e-tailer’s inventory allocation decision when customers are strategic. n MC25 North Bldg 131C Best Student Paper Competition - Session I Sponsored: Service Science Sponsored Session Chair: Aly Megahed, IBM Research - Almaden, San Jose, CA, 95123, United States 1 - Service Science Best Student Paper Competition Aly Megahed, IBM Research - Almaden, San Jose, CA, 95123, United States This session consists of finalists presentations (judged by an expert panel) to determine the Best Student Paper Award for the Service Science Cluster. n MC26 North Bldg 132A Service Consumer Sponsored: Service Science Sponsored Session Chair: Caner Canyakmaz, Ko University, Engineering Faculty, ENG 218, Istanbul / Sariyer, 34450, Turkey 1 - Tier Pricing Optimization the Revolving Fund Problem Ruben A. Proano, Associate Professor, Rochester Institute of Technology, Kate Gleason College of Engineering, 81 Lomb Memorial Drive, Rochester, NY, 14623, United States, Galo Eduardo Mosquera We use a three-stage optimization process to study alternative configurations of the Pan-American Health Organization Revolving Fund (RF), which is a single- tier group procurement mechanism used to buy vaccines for countries in the Americas. We propose an innovative approach to determine willingness-to-pay among vaccine buyers, and study the optimal number of tier prices that maximize affordability and profit, and the potential impact of large countries in the continent procuring vaccines independently. 2 - Two-dimensional Extended Warranties for Price-sensitive Customers Amitava Mitra, Professor, Auburn University, College of Business, Lowder Building Suite 419, Auburn, AL, 36849-5266, United States Products, such as, automobiles, have warranty policies that incorporate two dimensions, namely time since purchase of the product and usage based on accumulated mileage. While an initial warranty is offered during purchase of the product, options exist to renew the warranty in the event of no failure during the initial warranty. Here we consider extended warranty policies where the propensity to extend the warranty is influenced by the product price, relative ti a threshold value. The decision variables include the warranty time and usage limit in the original and extended policies, the product price, and the premium to be charged for extending the warranty. 3 - Improving Affordability in a Coordinated Non-cooperative Vaccine Market through Configuration of Group Buying Membership Bruno Alves Maciel, Rochester Institute of Technology, 59-6 Colony Manor Drive, Rochester, NY, 14623, United States, Ruben Proano, Galo Eduardo Mosquera We discuss the effect of allocating different countries into multiple vaccine purchasing groups whose procurement decisions are optimized by coordinating entities. These entities maximize savings for group members while offering profitability to vaccine producers. The vaccine purchasing groups form a hypothetical non-cooperative coordinated vaccine market. Coordinating entities choose their procurement plans via a two-stage optimization problem optimizing procurement decisions and pricing strategies. This study quanifies how group composition affects optimal procurement policies that incentivize higher affordability and profit in the vaccine market.

193