Online Model-Free Safety Verification for Markov Decision Processes Without Safety Violation.
Abhijit MazumdarRafal WisniewskiManuela-Luminita BujorianuPublished in: ECC (2024)
Keyphrases
- markov decision processes
- model free
- policy iteration
- reinforcement learning
- reinforcement learning algorithms
- risk sensitive
- average reward
- policy evaluation
- optimal policy
- function approximation
- transition matrices
- finite state
- state space
- infinite horizon
- temporal difference
- reachability analysis
- decision theoretic planning
- partially observable
- action space
- average cost
- least squares
- markov decision process
- planning under uncertainty
- data mining
- fixed point
- action sets
- model based reinforcement learning
- learning algorithm