Bias optimality for multichain continuous-time Markov decision processes.
Xianping GuoXinYuan SongJunyu ZhangPublished in: Oper. Res. Lett. (2009)
Keyphrases
- markov decision processes
- average reward
- stationary policies
- state space
- average cost
- optimal policy
- markov chain
- finite state
- policy iteration
- reinforcement learning
- action sets
- optimality criterion
- semi markov decision processes
- dynamic programming
- transition matrices
- reachability analysis
- planning under uncertainty
- finite horizon
- markov decision process
- reinforcement learning algorithms
- discounted reward
- optimal control
- partially observable
- dynamical systems
- state and action spaces
- factored mdps
- decision theoretic planning
- decision processes
- risk sensitive
- action space
- model based reinforcement learning
- decision problems
- optimal solution
- search space
- machine learning
- reward function