Fast probabilistic plan recognition without plan library
Minghao Yin, Hua-Jie Ou, Wen-Xiang Gu, Yin-Hua Lu, Ri-Xian Liu · 2005
Instead of using a plan library, the recognizer introduced in this paper uses a compact structure called PGoalGraph to represent goals, actions and states of the world. This method doesn't suffer the problem of acquisition and hand-coding a larger plan library as traditional methods do. The recognizer also extends classical methods in two directions. First, it doesn't make the assumption that all the effects of an action are deterministic. Second, the recognizer can produce an optimal plan instead of producing a set of equally plausible plan hypothesis. Since the recognizer is a domain independent one, the method can be adapted in almost every domain.