Compositional Bisimulation Minimization for Interval Markov Decision Processes.
Vahid HashemiHolger HermannsLei SongK. SubramaniAndrea TurriniPiotr J. WojciechowskiPublished in: LATA (2016)
Keyphrases
- markov decision processes
- finite state
- optimal policy
- state space
- dynamic programming
- reinforcement learning
- reachability analysis
- policy iteration
- transition matrices
- reinforcement learning algorithms
- planning under uncertainty
- model based reinforcement learning
- decision theoretic planning
- decision processes
- objective function
- average reward
- action space
- markov decision process
- risk sensitive
- partially observable
- reward function
- finite horizon
- factored mdps
- state and action spaces
- multistage
- average cost
- decision problems