Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WC35

4 - A Branch-and-price Algorithm for the Hitchcock-Koopmans Problem with Reusable Transportation Assets Tulay Flamand, Assistant Professor, Colorado School of Mines, Division of Economics and Business, Engineering Hall 816 15th Street, Room 313, Golden, CO, 80401, United States, Mohamed Haouari, Ghaith Rabadi We address a new variant of the Hitchcock-Koopmans problem that arises in commercial logistics where a given set of commodities must be transported from multiple supply nodes to multiple demand nodes using different types of transportation modes. Each transportation asset has a finite capacity and can be used for multiple consecutive rotations (i.e. round-trips). The problem requires delivering all commodities within a specified deadline while minimizing the total transportation cost. We describe an exact branch-and-price algorithm and present some preliminary computational experiments. 5 - Nudging Commuters’ Travel Choices with Dynamic and Personalized Incentives for Transportation Network Efficiency Yanshuo Sun, Florida State University, Tallahassee, FL, United States, Lei Zhang This paper explores an emerging strategy of transportation demand management, where dynamic and personalized incentives are transferred to travelers to influence their travel choices toward a more socially desirable state. A day-to-day traffic dynamics model is adopted, where commuters update their perceived travel time through trying and learning, based on which choices are made. A mathematical programming model is proposed for the system operator to design individually customized incentives daily. Through numerical analyses, we identify a paradox that increasing incentive limits do not necessarily improve the system efficiency due to probabilistic responses of travelers. Chair: Wei Wei, Waterloo University, Waterloo, ON, Canada 1 - Differentiable Linear Tracing Procedures for Selecting Nash Equilibrium and its Refinements Chuangyin Dang, Professor & Acting Head, City University of Hong Kong, 83 Tat Chee Avenue, Kowloon, Hong Kong The selection of a unique Nash equilibrium and its refinements has always been a concern in game theory. Nevertheless, Harsanyi’s linear tracing procedure can only select a unique Nash equilibrium for generic games. With the introduction of a differentiable quadratic function of an extra variable, this paper develops differentiable linear tracing procedures by incorporating concave quadratic terms into payoff functions. It is shown that the procedures are able to select, for every game, a unique Nash equilibrium, perfect equilibrium, proper equilibrium and perfect d-proper equilibrium, respectively. Numerical results further confirm the effectiveness and efficiency of the procedures. 2 - A Game Theory Model for Resource Allocation in Large Systems Design Soodabeh Yazdani, George Mason University, Takoma Park, MD, 20912, United States, Edward Huang Engineering design is a concurrent process with multiple engineers designing different components of a system simultaneously under the supervision of a project manager. Project manager allocates limited resources needed for the design process among the different engineers trying to optimize the overall value of the whole system. However, each team has its private information about the characteristics of the component being designed and the goal of each team is usually misaligned to the objective of the project manager. The goal of this paper is to develop an efficient mechanism for allocating resources in large systems design. 3 - Dynamic Decision Making with Asymmetric Information and Dependent States Deepanshu Vasal, University of Texas, Austin, TX, United States We consider finite horizon dynamic game with asymmetric information with N cooperative and selfish players where there exists an underlying state of the system that is a controlled Markov process, controlled by players’ actions. In each period, a player makes a common observation of the state and a private observation, and gets an instantaneous reward which is a function of the state and everyone’s actions. We present a sequential decomposition methodology to compute structured perfect Bayesian equilibria (SPBE) of this game, which computes SPBE in linear time. These equilibria exhibit signaling i.e. players’ actions reveal part of their private information that is payoff relevant to other players. n WC34 North Bldg 223 Game Theory II Contributed Session

4 - Modeling a Biofuel Supply Chain as an Equilibrium Problem with Equilibrium Constraints (EPEC): A Policy Analysis Saman Hamzeh, PhD Candidate, University of Wisconsin- Milwaukee, 3200 N. Cramer, Milwaukee, WI, 53211, United States We consider a biofuel supply chain problem in which a farmer supplies downstream refineries with non-identical crops. The problem has been modeled as a multi-leader-single-follower game to derive the farmer’s (follower) decisions on land use as well as refineries’ (leaders) proposed prices to the farmer. We join the leaders’ problem using EPEC formulation to find the NE. Parametric analyses are used to study the effect of subsidies on the profit of the players and the total social welfare of the supply chain. We observe that a government’s expenditure can be offset by the increase of the social welfare under certain circumstances. 5 - On the Smooth Pasting Principle and the Non-existence of Equilibrium Stopping Rules in Time Inconsistent Problems Wei Wei, Postdoctoral Fellow, Waterloo University, Waterloo, ON, Canada, Xunyu Zhou, Ken Seng Tan This paper considers a time inconsistent stopping problem. We show that properties of standard optimal stopping models may not be inherited when time consistency is lost. First, we focus on smooth pasting (SP). We find that the SP principle solves the time inconsistent problem within the intra-personal game theoretic framework if and only if a certain inequality is satisfied. Second, we formally show that the time inconsistent stopping problem does not admit any intra-personal equilibrium whenever the SP principle fails. The non-existence result and the failure of the SP principle caution blindly applying the SP principle to optimal stopping problems that are time inconsistent. n WC35 North Bldg 224A Aviation Economics Sponsored: Aviation Applications Sponsored Session Chair: Farbod Farhadi, Roger Williams University, Bristol, RI, 02809, United States 1 - Application of Random Utility Models for Airline Preference Estimation in Complex Operational Environment Ivan Tereshchenko, University of California, Berkeley, CA, United States Random utility models can be successfully applied to analyze decision-making strategies in settings when all decisions are made independently of each other. However, in air transportation, this is rarely the case. For example, aircraft operators make decisions about individual aircraft based on decisions made about all flights at their disposal. We use simulated schedules and optimization-based decision-making strategies to assess how well the random utility models can estimate parameters of airline objective functions in the realistic operational environment. 2 - Two-stage Optimized Recommendation Model: An Empirical Study of Airline Marketing on Website Visits Chen Zhang, UNSW, Sydney, Australia Traditional recommendation system utilises users’ historical purchases or browsing data to recommend products based on the similarity analysis, ignoring individual characteristics of the user and dynamic characteristics of the market. In this paper, we proposes a new two-stage recommendation model using optimisation theory. The first stage uses AI algorithm to extract each individual user’s preference feature. The second phase dynamically identifies users online intentions, dynamically adjusts recommended content, and provides personalised and optimised recommendations. The proposed model works better than the traditional recommendation model in the experiment. 3 - A Data-Driven Analysis of Equilibrium in the U.S. Airline Market Soheil Sibdari, University of Massachusetts, Brighton, MA, 02135, United States, Farbod Farhadi, David F. Pyke When multiple firms exist in a market, game theory can be used to analytically model their interactions. In such analysis, the optimal solution is replaced by an equilibrium. In this paper, we address an equilibrium analysis in the U.S. airline market using a large transactional data set. We also characterize the equilibrium outcome of the game and address its existence and uniqueness using the data set. We provide conditions under which multiple equilibria can be achieved.

481