NCSA staff who would like to submit an item for the calendar can email newsdesk@ncsa.illinois.edu.
Theory Seminar Series: Dr. Bhaskar Ray-Chaudhury, "Response-Aware Rational Equilibrium (RARE) in Pricing Games."

- Sponsor
- Siebel School of Computing and Data Science
- Speaker
- Dr. Bhaskar Ray Chaudhury
- Contact
- Chandra Chekuri
- chekuri@illinois.edu
- Views
- 17
- Originating Calendar
- Siebel School Speakers Calendar
Abstract: We study markets in which sellers set prices to maximize revenue and buyers choose their most preferred bundles within fixed budgets. Such budgets naturally model delegated purchasing, where an AI agent has a spending limit authorized by its user. Our model covers both goods with limited supply and non-rival goods, such as data, that can be sold to many buyers. We show that pure Nash equilibria may fail to exist even with simple buyer preferences: there may be no prices at which every seller prefers to keep its price unchanged, assuming competitors keep theirs unchanged. However, some seemingly profitable price changes lose their appeal once competitors respond in their own interests. Motivated by this observation, we introduce response-aware rational equilibrium (RARE). Competitors may respond jointly to a seller’s price change, but every competitor must earn at least as much under their joint response as it earned immediately after the initial change. Prices form a RARE if every profitable unilateral price change can be countered by such a response that eliminates the deviating seller’s gain. We study the existence and structure of RARE, as well as its emergence through natural market dynamics in which sellers adjust prices to improve their revenues. These questions are particularly relevant to AI sellers capable of anticipating competitors’ reactions, making the effects of strategic reasoning on market stability a concrete question of agent design.
Bio: Bhaskar Ray Chaudhury is an Assistant Professor in the Department of Industrial and Enterprise Systems Engineering at the University of Illinois Urbana-Champaign. He led the IDEAL Lab (Incentives, Data, Equilibria, Allocations, and Learning) at Illinois. He has worked extensively on economics of data, equilibrium computation, mechanism design, and theory of fair allocations. He is broadly interested in economics and computation, machine learning, and theoretical computer science.