Policy Improvement by a Model-Free Dyna Architecture.
Kao-Shing HwangChia-Yue LoPublished in: IEEE Trans. Neural Networks Learn. Syst. (2013)
Keyphrases
- model free
- rl algorithms
- function approximation
- policy iteration
- temporal difference
- reinforcement learning
- policy evaluation
- reinforcement learning algorithms
- function approximators
- average reward
- optimal policy
- temporal difference learning
- reinforcement learning methods
- learning algorithm
- markov decision problems
- evaluation function
- radial basis function
- least squares
- pattern recognition