Near-continuous time Reinforcement Learning for continuous state-action spaces.
Lorenzo CroissantMarc AbeilleBruno BouchardPublished in: CoRR (2023)
Keyphrases
- continuous state
- action space
- state space
- reinforcement learning
- stochastic processes
- markov decision processes
- optimal control
- policy search
- control policies
- real valued
- markov chain
- continuous state spaces
- optimal policy
- state action
- state dependent
- markov decision problems
- markov decision process
- continuous action
- dynamic programming
- heuristic search
- dynamical systems
- machine learning
- function approximators
- reinforcement learning algorithms
- action selection
- finite state
- robot navigation
- temporal difference
- planning problems
- random fields
- probability distribution
- model free
- search space
- function approximation
- partially observable markov decision processes
- random variables
- mobile robot
- multi agent