Title: Markov Persuasion Processes with Endogenous Agent Beliefs
Venue: IEOR Seminar Room
Date and time: 3 January 2025 (Friday), 10:00 – 11:00 a.m.
Abstract: We study a Markov persuasion process, where a long-lived principal (``sender'') persuades a stream of short-lived agents by sharing information about an evolving state. The state transitions are Markovian and the sender seeks to maximize the long-run average reward by committing to a (possibly history-dependent) signaling mechanism. While most previous studies of Markov persuasion consider agents with exogenously specified beliefs, we study a more natural variant with endogenous beliefs that depend on the chain's realized history. A key challenge in studying this variant is to model the agents' partial knowledge about the history. We focus on settings where each agent observes the history with a $k$ period lag. We show that the sender's long-run average payoff (weakly) increases with the lag, and provide efficient LP formulations to compute the optimal signaling mechanisms for $k = 0$ (full history information) and $k = \infty$ (no history information). For general values of the lag, we identify challenges in formulating the sender's problem, and provide a bilinear optimization framework to solve a restricted problem. Finally, using robust persuasion, we design simple signaling mechanisms that are approximately optimal when the lag is large.
Joint work with You Zu and Haifeng Xu
Bio: Krishnamurthy Iyer is an Associate Professor at the Department of Industrial and Systems Engineering (ISyE) at the University of Minnesota. His research interests lie in the design and analysis of markets and service systems, with focus on methodological research at the intersection of operations research, economics and computation. He obtained his PhD in Management Science and Engineering at Stanford University, where he received a Stanford Graduate Fellowship and a Dantzig-Lieberman Fellowship. He serves as an Associate Editor at Management Science, and regularly serves on the program committee for the ACM Conference on Economics and Computation (EC). He is the recipient of the First Place in the 2017 INFORMS Junior Faculty Interest Group (JFIG) Paper Competition, and a College of Engineering Excellence in Teaching Award from Cornell.