2025/26 eat & LEARN Seminar Series (2)

• Title: Algorithmic Game Theory: Introduction and Recent Progress
• Speaker: Suho Shin (Department of Computer Science, University of Maryland)
• Venue: Room #302, Science Bldg. IV
• Host: Prof. Kangbok Lee (Dept. of Industrial and Management Engineering)

Abstract

Algorithmic Game Theory (AGT)—broadly referred to as Economics and Computation (EconCS)—studies the design of algorithms when the inputs of problem instances are provided by multiple strategic agents, such as (ir)rational human participants and goal-oriented AI agents. The field has received rapidly growing attention over the past few decades due to its substantial impact on online advertising auctions, data markets, and modern online platforms. With the rise of multi-agent paradigms in the era of large language models (LLMs), AGT has become increasingly central to computer science, economics, operations research, and a broader set of engineering and social science communities.

The core thread of AGT seeks to advance game theory through an algorithmic lens—for example, by analyzing the computational hardness of equilibrium concepts, overcoming classical impossibility results in theoretical economics with approximate yet efficient outcomes, and designing scalable algorithms for markets and organizations. Such applications include bidding algorithms and auction mechanisms for online advertisement auctions, matching algorithms for kidney exchange and school choice, and algorithms for fair division of goods and chores.

In this talk, I will present seminal progress in AGT with historical context, and discuss recent research directions, open problems/remaining challenges, and finally my recent research on principal-agent problems in this context.