Dyna-style Model-based reinforcement learning with Model-Free Policy Optimization.
Kun DongYongle LuoYuxin WangYu LiuChengeng QuQiang ZhangErkang ChengZhiyong SunBo SongPublished in: Knowl. Based Syst. (2024)
Keyphrases
- model free
- model based reinforcement learning
- rl algorithms
- reinforcement learning
- policy iteration
- function approximation
- markov decision processes
- temporal difference
- policy evaluation
- optimal policy
- reinforcement learning algorithms
- function approximators
- average reward
- temporal difference learning
- markov decision problems
- reinforcement learning methods
- state space
- markov decision process
- learning problems
- reward function
- partially observable
- dynamic programming
- neural network
- decision processes
- action space
- action selection
- infinite horizon
- monte carlo