基于U-规划图与动作效果相关的规划识别研究

Research of Planning Recognition Based on U-Planning Graph and Related with Action Effect

针对支持动作效果是多个的情况,智能体如何根据现有的信息为该效果确定最终的支持动作的问题进行了研究,并给出了相应的FSas算法。FSas算法分别从命题层和动作层出发,利用被观察到的动作效果或动作以及互斥计算减小了支持动作的空间,从而求得最终支持效果的动作,实现了规划识别。研究结果表明,该算法不仅解决了支持动作效果的动作是唯一的情况,而且进一步解决了支持动作是多个的情况。同时以实例说明了该算法的执行过程,表明该算法是合理可行的。

We do a reserch on the reasoning of action effects that observer observed in present planning recognition system. It puts forward the FSas algorithm, this algorithm solves the case that action supporting action effects is unique, and solves the case that supporting actions are many. The FSas algorithm starts from the proposition layer and action layer and uses the observed action effects, actions and mutually exclusive computing to reduce the space of supported actions, which can determine the actions that support the action effect and further realize planning recognition. Example shows that the of algorithm＇s execution process and proves that the algorithm is reasonable and feasible.