Verification of General Markov Decision Processes by Approximate Similarity Relations and Policy Refinement.
Sofie HaesaertSadegh Esmaeil Zadeh SoudjaniAlessandro AbatePublished in: SIAM J. Control. Optim. (2017)
Keyphrases
- markov decision processes
- optimal policy
- policy evaluation
- policy iteration
- markov decision process
- finite horizon
- factored mdps
- infinite horizon
- state space
- reinforcement learning
- partially observable
- finite state
- average cost
- average reward
- decision processes
- dynamic programming
- action space
- similarity relations
- state and action spaces
- transition matrices
- decision theoretic planning
- reward function
- total reward
- reinforcement learning algorithms
- least squares
- partially observable markov decision processes
- neural network
- temporal difference
- long run
- markov decision problems
- decision problems
- multistage
- fixed point
- model checking
- sufficient conditions
- expected reward
- linear programming
- action sets
- machine learning