Near-Optimal Randomized Exploration for Tabular Markov Decision Processes.
Zhihan XiongRuoqi ShenQiwen CuiMaryam FazelSimon S. DuPublished in: NeurIPS (2022)
Keyphrases
- markov decision processes
- model based reinforcement learning
- interval estimation
- state space
- optimal policy
- dynamic programming
- reinforcement learning
- transition matrices
- finite state
- policy iteration
- factored mdps
- reinforcement learning algorithms
- reachability analysis
- average reward
- decision theoretic planning
- average cost
- action space
- planning under uncertainty
- action sets
- infinite horizon
- decision processes
- state abstraction
- finite horizon
- semi markov decision processes
- state and action spaces
- markov decision process
- partially observable