A Novel Prioritization Technique for Solving Markov Decision Processes.
Jilles Steeve DibangoyeBrahim Chaib-draaAbdel-Illah MouaddibPublished in: FLAIRS Conference (2008)
Keyphrases
- markov decision processes
- transition matrices
- semi markov decision processes
- state space
- optimal policy
- reinforcement learning
- policy iteration
- dynamic programming
- finite state
- markov decision problems
- reinforcement learning algorithms
- stochastic shortest path
- planning under uncertainty
- decision processes
- partially observable
- finite horizon
- decision theoretic planning
- average cost
- average reward
- policy iteration algorithm
- model based reinforcement learning
- factored mdps
- reachability analysis
- infinite horizon
- action space
- action sets
- state and action spaces
- risk sensitive
- linear program
- reward function
- machine learning