Metrics for Markov Decision Processes with Infinite State Spaces
Norman FernsPrakash PanangadenDoina PrecupPublished in: CoRR (2012)
Keyphrases
- markov decision processes
- state space
- reinforcement learning algorithms
- reinforcement learning
- finite state
- optimal policy
- transition matrices
- policy iteration
- dynamic programming
- markov chain
- decision theoretic planning
- planning under uncertainty
- model based reinforcement learning
- partially observable
- factored mdps
- state abstraction
- reachability analysis
- markov decision process
- action space
- finite horizon
- reward function
- infinite horizon
- heuristic search
- average cost
- planning problems
- search space
- average reward
- action sets
- markov decision problems
- continuous state spaces
- finite number
- discounted reward