Robust Distributed Nash Equilibrium Seeking of Multi-Agent Games Against Cyber-Attacks

This paper addresses the challenges posed by cyber-attacks in multi-agent games, in which the interactive information among agents may be maliciously tampered and agents lack the means to verify the accuracy of received information. To solve this problem, we introduce the concepts of a trustworthy graph and a robust information graph. Utilizing these concepts, we propose a robust distributed algorithm to accurately estimate global information by collecting and filtering extreme values, ultimately converging to a Nash equilibrium. Our research demonstrates that, under specific graph topology conditions, the algorithm exhibits robustness against cyber-attacks leading to data intrusion.