Region-based value iteration for partially observable Markov decision processes.
Hui LiXuejun LiaoLawrence CarinPublished in: ICML (2006)
Keyphrases
- partially observable markov decision processes
- finite state
- decision problems
- dynamical systems
- reinforcement learning
- partially observable markov
- markov decision processes
- belief state
- optimal policy
- dynamic programming
- planning under uncertainty
- continuous state
- belief space
- state space
- planning problems
- multi agent
- image segmentation
- partially observable stochastic games
- stochastic domains
- partially observable
- partially observable domains
- average reward
- infinite horizon
- sequential decision making problems
- approximate solutions
- markov chain
- state variables
- belief revision
- predictive state representations
- special case
- model checking