Multi-Agent Reinforcement Learning with General Information Structures: Convergence to Equilibria
摘要
So far we have studied in Chaps. 21 and 22 learning in single-agent models with continuous state and action spaces, in fully observed as well as partially observed settings. In this chapter, we move on to multi-agent systems and discuss learning theoretic methods for decentralized information structure models, for both stochastic teams and games.