Efficient and Safe Exploration in Deterministic Markov Decision Processes with Unknown Transition Models.
Erdem BiyikJonathan MargoliashShahrouz Ryan AlimoDorsa SadighPublished in: ACC (2019)
Keyphrases
- markov decision processes
- transition matrices
- reinforcement learning
- model based reinforcement learning
- finite state
- optimal policy
- finite horizon
- dynamic programming
- state space
- reachability analysis
- action sets
- probabilistic model
- decision processes
- interval estimation
- factored mdps
- policy iteration
- decision theoretic planning
- average reward
- reinforcement learning algorithms
- search space