Threshold Constraints with Guarantees for Parity Objectives in Markov Decision Processes.
Raphaël BerthonMickael RandourJean-François RaskinPublished in: CoRR (2017)
Keyphrases
- markov decision processes
- state space
- optimal policy
- reinforcement learning
- transition matrices
- finite state
- dynamic programming
- decision theoretic planning
- average cost
- action space
- average reward
- policy iteration
- decision processes
- finite horizon
- reinforcement learning algorithms
- planning under uncertainty
- reachability analysis
- action sets
- multiple objectives
- partially observable
- risk sensitive
- infinite horizon
- continuous state spaces
- semi markov decision processes
- state abstraction
- reward function
- state and action spaces
- initial state
- probabilistic planning
- real time dynamic programming