Automatic Discovery of Subgoals in Reinforcement Learning Using Strongly Connected Components.
Seyed Jalal KazemitabarHamid BeigyPublished in: ICONIP (1) (2008)
Keyphrases
- automatic discovery
- strongly connected components
- reinforcement learning
- shortest path
- connected components
- web services
- semantic web services
- state space
- argumentation semantics
- machine learning
- closed contours
- multi agent
- learning process
- dynamic programming
- optimal policy
- max sat
- variable ordering
- computer vision
- np complete