Iterative Policy-Space Expansion in Reinforcement Learning.
Jan Malte LichtenbergÖzgür SimsekPublished in: CoRR (2019)
Keyphrases
- reinforcement learning
- optimal policy
- action space
- policy search
- markov decision process
- action selection
- policy iteration
- function approximators
- control policy
- reinforcement learning algorithms
- infinite horizon
- function approximation
- space time
- partially observable domains
- approximate dynamic programming
- actor critic
- partially observable environments
- higher dimensional
- optimal control
- markov decision processes
- real robot
- control problems
- finite state
- rl algorithms
- control policies
- low dimensional
- state space
- search space
- state and action spaces
- learning algorithm
- neural network