Efficient reinforcement learning in continuous state and action spaces with Dyna and policy approximation.
Shan ZhongQuan LiuZongzhang ZhangQiming FuPublished in: Frontiers Comput. Sci. (2019)
Keyphrases
- reinforcement learning
- continuous state and action spaces
- function approximators
- optimal policy
- function approximation
- approximation methods
- action selection
- continuous state
- state space
- policy search
- multi agent
- rl algorithms
- temporal difference
- control policies
- temporal difference learning
- markov decision process
- neural network
- machine learning
- action space
- reinforcement learning algorithms
- control policy
- markov decision processes
- domain independent
- decision making
- learning algorithm