Robust Synchronization in Markov Decision Processes.
Laurent DoyenThierry MassartMahsa ShirmohammadiPublished in: CONCUR (2014)
Keyphrases
- markov decision processes
- optimal policy
- finite state
- state space
- reinforcement learning
- transition matrices
- factored mdps
- decision theoretic planning
- partially observable
- policy iteration
- reinforcement learning algorithms
- finite horizon
- reachability analysis
- decision processes
- markov decision process
- dynamic programming
- action space
- planning under uncertainty
- average cost
- average reward
- infinite horizon
- model based reinforcement learning
- state abstraction
- reward function
- expected reward
- action sets