Limit Synchronization in Markov Decision Processes.
Laurent DoyenThierry MassartMahsa ShirmohammadiPublished in: FoSSaCS (2014)
Keyphrases
- markov decision processes
- finite state
- state space
- policy iteration
- optimal policy
- dynamic programming
- factored mdps
- transition matrices
- reinforcement learning
- reinforcement learning algorithms
- average cost
- decision theoretic planning
- risk sensitive
- model based reinforcement learning
- average reward
- markov decision process
- finite horizon
- infinite horizon
- action sets
- decision processes
- reachability analysis
- action space
- state and action spaces
- discounted reward
- policy evaluation
- semi markov decision processes
- reward function
- markov chain
- interval estimation