Dealing with Non-Stationarity in MARL via Trust-Region Decomposition.
Wenhao LiXiangfeng WangBo JinJunjie ShengHongyuan ZhaPublished in: ICLR (2022)
Keyphrases
- trust region
- multi agent reinforcement learning
- global optimum
- column generation
- optimization methods
- hessian matrix
- log likelihood
- newton method
- line search
- multi agent
- mean shift
- learning agents
- optimization method
- multi agent learning
- multi agent systems
- information theoretic
- linear programming
- neural network
- levenberg marquardt