Reinforcement Learning in Continuous Action Spaces through Sequential Monte Carlo Methods.
Alessandro LazaricMarcello RestelliAndrea BonariniPublished in: NIPS (2007)
Keyphrases
- action space
- reinforcement learning
- state space
- markov decision processes
- sequential monte carlo methods
- state and action spaces
- real valued
- continuous state
- continuous state spaces
- stochastic processes
- state action
- function approximators
- action selection
- function approximation
- learning algorithm
- continuous action
- markov decision process
- model free
- reinforcement learning algorithms
- dynamic programming
- gaussian model
- optimal policy
- mobile robot
- single agent
- search space
- image intensity
- prior knowledge
- supervised learning
- skill learning