Continuous simulation abstraction refinement for Markov decision processes.
Xu GuoZongyuan YangPublished in: ICSAI (2017)
Keyphrases
- markov decision processes
- decision theoretic planning
- action space
- state space
- state abstraction
- optimal policy
- dynamic programming
- transition matrices
- reinforcement learning
- finite state
- policy iteration
- planning under uncertainty
- average cost
- reinforcement learning algorithms
- markov decision process
- reachability analysis
- continuous state spaces
- infinite horizon
- partially observable
- factored mdps
- finite horizon
- macro actions
- average reward
- decision processes
- state and action spaces
- risk sensitive
- model based reinforcement learning
- action sets
- search algorithm
- multi agent
- optimal control
- optimal solution
- probabilistic planning
- heuristic search
- model checking
- real valued