Threshold Constraints with Guarantees for Parity Objectives in Markov Decision Processes.
Raphaël BerthonMickael RandourJean-François RaskinPublished in: ICALP (2017)
Keyphrases
- markov decision processes
- optimal policy
- state space
- finite state
- dynamic programming
- transition matrices
- policy iteration
- reinforcement learning
- decision theoretic planning
- average reward
- reachability analysis
- reinforcement learning algorithms
- partially observable
- finite horizon
- model based reinforcement learning
- factored mdps
- planning under uncertainty
- average cost
- action space
- reward function
- decision processes
- action sets
- markov decision process
- data mining
- state abstraction
- infinite horizon
- multiple objectives
- state and action spaces
- sufficient conditions
- least squares
- machine learning