Research on Algorithms for Differential Graphical Game and Distributed Nash Equilibrium in Continuous Systems
Kuankuan Qi, Fuxiao Tan · 2023
In recent years, game theory, graph theory, multi-agent learning methods, and multi-agent control have developed from single-discipline to multidisciplinary. The attention to deep interdisciplinary integration is becoming increasingly apparent. This interdisciplinary integration has achieved corresponding research results in fields such as artificial intelligence and automatic control, and has become a current research hotspot. To timely reflect this academic dynamic and trend, this paper focuses on the differential graphical game and distributed Nash equilibrium in continuous systems. First, normal-form games and differential graphical games were introduced. Second, multi-agent reinforcement learning methods based on policy iteration and value iteration were studied to solve three types of stochastic game problems and the distributed Nash equilibrium problem in continuous systems was focused on discussion. Then, recent research progress on differential graphical games and distributed Nash equilibrium in continuous systems were reviewed. Finally, a summary and outlook were provided for this cross-disciplinary area.