Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MD23

2 - When Prospect Theory Meets Consumer Choice Models: The Role of Reference Prices Ruxian Wang, Johns Hopkins University, Carey Business School, 100 International Dr, Baltimore, MD, 21202, United States Reference prices arise as price expectations against which consumers evaluate products in their purchase scenarios. We investigate what will happen when prospect theory (e.g., reference prices) meets consumer choice models from the perspectives of both the consumers and the firm. 3 - Optimal Information Control Strategies for the Online P2P Platform Shu Hu, Ningbo Supply Chain Innovation Institute China, Shiliang Cui, Mingcheng Wei We investigate the optimal information control strategies for the online P2P platform, which will decide whether to provide product information for the buyer, and whether to provide market information for the seller, when it aims to maximize the average transaction volume and the average transaction amount, respectively. We further examine the value of product information for buyers and the value of market information for sellers. 4 - Dynamic Pricing of the Fixed-term Subscription Contracts or One-time Purchase Offered to the Strategic Customers in a Heterogeneous Market Roozbeh Yousefi, Queen’s University, 310 Bath Rd., Unit 1112, Kingston, ON, K7M9H1, Canada, Jue Wang, Yuri Levin, Mikhail Nediak Subscriptions are agreements between a company which commits to deliver a service or provide access to a service and its customers. We present a continuous time dynamic pricing model for a monopolist offering a fixed term subscription contract without per-use charges and limit of access or a one-time purchase at all the instances, to strategic customers in a growing heterogeneous market. We modeled the discrete choice problem of customers and used it to formulate the monopolist’s problem in terms of optimal control, derive its optimality conditions. We demonstrate the optimal pricing results in numerical experiments. 5 - Online Decision-making with High-dimensional Covariates and Binary Response Wang Xue, Penn State University, Leonhard, Building, 310 S. Barnard St, University Park, PA, United States, Mingcheng Wei, Tao Yao In this study, we consider the online learning and decision-making problem with high-dimensional covariates and binary response. We penalize the bandit model with Minimax Concave Penalty (MCP-bandit) to handle the high-dimensional issue. Under epsilon-decay sampling scheme, we show the cumulative regret of MCP-bandit is upper bound O(s^3log d log T), where s is the sparsity level, d is the total dimension and T is the time length. Moreover, we prove the log T dependence of cumulative regret is optimal. Sponsored: Finance Sponsored Session Chair: Xin Guo, University of California-Berkeley, Piedmont, CA, 94611, United States 1 - A Stochastic Game and Moving Free Boundary Problem Renyuan Xu, University of California-Berkeley, 4141 Etcheverry Hall, OR Department, Berkeley, CA, 94720, United States, Xin Guo, Wenpin Tang Stochastic control problems are closely related to free boundary problems, where both the underlying fully nonlinear PDEs and the separating boundaries are integral parts of the problems. In this talk, we propose a class of stochastic games and show how the free boundary problems involve moving boundaries due to the additional game nature. We will provide explicit solutions in terms of Nash equilibria by solving a Skorokhod problem with moving boundaries. We will use some special cases of the games in light of the classical finite fuel problem to compare game strategies in terms of pooling and sharing. We will also discuss the Nash equilibrium strategies in the framework of controlled ranked SDEs. 2 - A Stochastic Numerical Method for Mean Field Games Mathieu Lauriere, Princeton University, Princeton, NJ, United States We present a new stochastic algorithm to solve mean field games and optimal control problems of McKean-Vlasov dynamics. This numerical method relies on the system of forward-backward stochastic differential equations characterizing the solutions to these problems. Several examples of applications will be provided. This is joint work with Ren Carmona. n MD23 North Bldg 131A Joint Session FSS/Practice Curated: Stochastic Games with Applications

3 - The Coordination of Centralized and Distributed Generation Matteo Basei, University of California, Berkeley, Berkeley, CA, United States, Ren Aid, Huyen Pham We analyse the interaction between centralised carbon-emissive technologies and distributed non-emissive technologies. A representative consumer can satisfy her electricity demand by investing in solar panels and by buying power from a centralised firm. We consider the point of view of the consumer, the firm and a social planner, formulating suitable McKean-Vlasov control problems with stochastic coefficients. First, we provide explicit formulas for the production strategies which minimise the costs. Then, we look for an equilibrium price. 4 - Non-zero Sum Stochastic Games with Impulse Control Haoyang Cao, University of California, Berkeley, Berkeley, CA, United States This is a joint work with Prof. Xin Guo and Mr. Matteo Basei. We generalize the single-agent impulse control problem, i.e. the cash management problem as in [Constantinides and Richard, 1978], to an N-player impulse game. In this impulse game, each agent controls its own state dynamics and they are coupled together through cost functions. We provide two versions of verication theorems, one requiring value functions to be continuous everywhere, the other relaxing the regularity condition and adding assumptions on cost structures instead. Under a symmetric setting, we are able to give a semi-explicit Nash equilibrium. n MD24 North Bldg 131B E-Business Sponsored: EBusiness Sponsored Session Chair: Cheng Nie, TX, United States 1 - Does Money Talk? The Evidence from an Online Peer to Peer Car Sharing Platform Yixin Lu, The George Washington University, 2201 G. Street NW, Funger Hall 506, Washington, DC, 20052, United States We conduct a randomized field experiment at a large online peer-to-peer car sharing platform to examine the impact of monetary incentives on user onboarding. We find that monetary incentives are no better than simple email reminders in encouraging self-disclosure of private information nor user engagement with the platform. 2 - Predicting Time to Upgrade Under Successive Product Generations Xinxue Qu, Iowa State University, 4815 Todd Drive, 71, Ames, IA, 50014, United States, Aslan Lotfi, Zhengrui Jiang In the presence of successive product generations, customers may decide to purchase a future generation before its release. As a result, sales of the new generation often show a declining pattern. In this study, we propose an exponential-decay survival model to predict customers’ time to upgrade decisions. Empirical analysis using data for a major sports video game series shows that the Expo-Decay model performs better in prediction accuracy. Empirical results reveal that previous adoption and usage habits impact upgrade decisions: (i) potential switching customers are more likely to upgrade; (ii) heavy players tend to upgrade earlier; (iii) specialized customers are less likely to upgrade. 3 - Predicting Stock Price Movements via Multi-relational Inter-firm Networks John Rios, University of Iowa, Iowa City, IA, United States, Kang Zhao, Nick Street Predicting stock movements is challenging, but has attracted tremendous amount of attention from both practitioners and researchers. At the same time, firms interact with each other in a multi-relational network with multiple types of relationships. Using real-world supply and competition networks among more than 1,000 firms, we predict a firm’s stock movements by leveraging performance of its customers, suppliers, and competitors. We show that features based on network neighbors of a firm significantly contribute to the prediction of its future stock movements. Additional analyses show that suppliers’ and competitors’ performance is more indicative of firms’ stock movement than customers’

222