Continuous Action Reinforcement Learning Automata - Performance and Convergence.
Abdel RodríguezRicardo Grau ÁbaloAnn NowéPublished in: ICAART (2) (2011)
Keyphrases
- continuous action
- policy search
- reinforcement learning
- continuous state
- finite state
- action space
- partially observable markov decision processes
- continuous state and action spaces
- markov decision processes
- reinforcement learning algorithms
- state space
- convergence speed
- convergence rate
- reward function
- function approximation
- robot navigation
- optimal policy
- model checking
- dynamic programming
- learning algorithm
- control strategies
- partially observable
- dynamic environments
- policy gradient