Faster Algorithms for Markov Decision Processes with Low Treewidth
Krishnendu ChatterjeeJakub LackiPublished in: CoRR (2013)
Keyphrases
- markov decision processes
- policy iteration
- factored mdps
- reachability analysis
- finite state
- learning algorithm
- dynamic programming
- reinforcement learning
- planning under uncertainty
- transition matrices
- space complexity
- infinite horizon
- state space
- risk sensitive
- state and action spaces
- interval estimation
- stochastic shortest path
- continuous state spaces
- finite horizon
- reinforcement learning algorithms
- np hard
- search space
- computational complexity