Guiding the search in continuous state-action spaces by learning an action sampling distribution from off-target samples.
Beomjoon KimLeslie Pack KaelblingTomás Lozano-PérezPublished in: CoRR (2017)
Keyphrases
- continuous state
- action space
- reinforcement learning
- continuous action
- state action
- state space
- action selection
- policy search
- markov decision processes
- finite state
- robot navigation
- learning algorithm
- real valued
- search algorithm
- continuous state spaces
- search space
- state dependent
- domain independent
- function approximation
- stochastic processes
- stochastic games
- control policies
- decision making