Querying to Find a Safe Policy under Uncertain Safety Constraints in Markov Decision Processes.
Shun ZhangEdmund H. DurfeeSatinder P. SinghPublished in: AAAI (2020)
Keyphrases
- markov decision processes
- optimal policy
- policy iteration
- markov decision process
- average reward
- finite horizon
- infinite horizon
- state space
- average cost
- state and action spaces
- reinforcement learning
- action space
- decision processes
- finite state
- partially observable
- dynamic programming
- reward function
- policy iteration algorithm
- policy evaluation
- decision problems
- total reward
- expected reward
- transition matrices
- decision theoretic planning
- long run
- decision making
- factored mdps
- planning under uncertainty
- reachability analysis
- discounted reward
- markov decision problems
- stationary policies
- action sets
- sufficient conditions
- control policies
- reinforcement learning algorithms
- model based reinforcement learning
- model free
- approximate dynamic programming
- decision diagrams
- action selection
- least squares
- multistage
- search space
- temporal difference