Distributional Soft Actor-Critic: Off-Policy Reinforcement Learning for Addressing Value Estimation Errors.
Jingliang DuanYang GuanShengbo Eben LiYangang RenQi SunBo ChengPublished in: IEEE Trans. Neural Networks Learn. Syst. (2022)
Keyphrases
- actor critic
- estimation error
- reinforcement learning
- temporal difference
- reinforcement learning algorithms
- policy gradient
- approximate dynamic programming
- step size
- optimal control
- neuro fuzzy
- gradient method
- policy iteration
- function approximation
- standard deviation
- error rate
- state space
- covariance matrix
- markov decision processes
- temporal difference learning
- dynamic programming
- model free
- average reward
- policy gradient methods
- rl algorithms
- learning algorithm
- cost function
- multi agent
- multiscale
- convergence speed
- convergence rate
- linear program
- dynamical systems
- monte carlo
- reinforcement learning methods
- particle swarm optimization
- objective function