多智能体强化学习

多个智能体的强化学习：合作与竞争设置、通信和多机器人协调。

Learning Objectives

1. From Single-Agent to Multi-Agent

2. Decentralized POMDP

3. Cooperative MARL

3.1 QMIX

3.2 MAPPO

3.3 Communication Protocols

4. Competitive MARL

4.1 Self-Play

4.2 League Training

5. Multi-Robot RL

5.1 Swarm Coordination

5.2 Heterogeneous Teams

6. MARL Frameworks (PettingZoo, EPyMARL)

Exercises

References