F2A2: Flexible Fully-decentralized Approximate Actor-critic for Cooperative Multi-agent Reinforcement Learning.
Wenhao LiBo JinXiangfeng WangJunchi YanHongyuan ZhaPublished in: CoRR (2020)
Keyphrases
- multi agent reinforcement learning
- cooperative
- actor critic
- reinforcement learning
- multi agent
- multi agent systems
- policy gradient
- temporal difference
- learning agents
- multi agent learning
- reinforcement learning algorithms
- optimal control
- gradient method
- function approximation
- neuro fuzzy
- distributed control
- stochastic games
- autonomous agents
- multiagent systems
- approximate solutions
- machine learning
- rl algorithms
- average reward
- game theory
- markov decision processes
- state space