Ph.D. Defense

Yue Guan

(Advisor: Prof. Panagiotis Tsiotras)

“Scalable and Provable Decision-Making for Large-Population 

Multi-Agent Systems in Complex Domains”

On 

Thursday, October 2

11:00 a.m.

Montgomery Knight 317

https://gatech.zoom.us/j/5829364964?omn=93052220854

Abstract

Modern autonomous systems often operate in large-scale teams and dynamic, uncertain environments, where traditional decision-making frameworks struggle with scalability, information asymmetry, and adversarial interactions. This thesis develops new game-theoretic and learning-based methods to address these challenges across three fronts: (i) large-population games for scalable coordination and competition among robot teams, (ii) hierarchical game formulations for efficient decision-making in large state spaces, and (iii) asymmetric-information games that capture intentional concealment of hierarchical representations in adversarial settings. The proposed methods provide provable guarantees—including existence of equilibria, convergence of learning algorithms, and performance guarantees and exploitability bounds—and are validated through applications in pursuit-evasion, adversarial resource allocation, and team defense–attack scenarios. Together, these contributions establish a principled foundation for robust and scalable decision-making in large, adversarial, and uncertain multi-agent environments.

Committee 

• Dr. Panagiotis Tsiotras – School of Aerospace Engineering (advisor)
• Dr. Kyriakos Vamvoudakis – School of Aerospace Engineering
• Dr. Matthew Gombolay – School of Interactive Computing
• Dr. Vidya Muthukumar – School of Electrical and Computer Engineering
• Dr. Vijay Kumar – Departments of Mechanical Engineering and Applied Mechanics, University of Pennsylvania