Exploiting Domain Symmetries in Reinforcement Learning with Continuous State and Action Spaces.
Alejandro AgostiniEnric CelayaPublished in: ICMLA (2009)
Keyphrases
- continuous state and action spaces
- reinforcement learning
- continuous state
- action selection
- function approximation
- domain independent
- function approximators
- domain specific
- learning algorithm
- action space
- state space
- dynamic programming
- model free
- policy search
- machine learning
- robot navigation
- markov decision process
- control system
- temporal difference learning
- search space