Combination of acceleration procedures for solving stochastic shortest-path Markov decision processes.
Ma. de Guadalupe García-HernándezJosé Ruiz-PinalesSergio E. Ledesma-OrozcoJuan Gabriel Aviña-CervantesEva OnaindiaAlberto Reyes-BallesterosPublished in: ISKE (2010)
Keyphrases
- stochastic shortest path
- markov decision processes
- markov decision problems
- finite state
- transition matrices
- optimal policy
- semi markov decision processes
- partially observable
- state space
- reinforcement learning
- average cost
- decision processes
- planning under uncertainty
- policy iteration
- reinforcement learning algorithms
- average reward
- decision theoretic planning
- infinite horizon
- multi agent