RiskQ: Risk-sensitive Multi-Agent Reinforcement Learning Value Factorization.
Siqi ShenChennan MaChao LiWeiquan LiuYongquan FuSongzhu MeiXinwang LiuCheng WangPublished in: CoRR (2023)
Keyphrases
- multi agent reinforcement learning
- risk sensitive
- optimal control
- reinforcement learning
- markov decision processes
- model free
- utility function
- stochastic games
- multi agent
- multi agent learning
- markov decision problems
- learning agents
- multi agent systems
- control policies
- optimality criterion
- decision theoretic
- optimal policy
- expected utility
- average cost
- cooperative
- reinforcement learning algorithms
- random walk
- decision makers
- machine learning
- multistage
- probability distribution
- dynamic programming
- bayesian networks