【摘要】 基于规划识别中的规划分层模型,利用单元素集树状网络和单连通网络两种算法,提出了一种求解规划识别问题的贝叶斯网络,该网络以图形的模式表示了规划识别中的因果信息,体现了代理的目的和行为之间的连接概率分布.并以一个实例说明了该网络中结点之间的信息传播过程,表明了使用该网络来求解规划识别问题的可行性.更多还原

【Abstract】 Based on the hierarchy model of plans in plan recognition, a Bayesian network for solving problems of plan recognition is presented by using the algorithm for singleton tree networks and the algorithm for single connected networks. The network is a graphical model for expressing the causal information in plan recognition and it represents the joint probabilistic distribution between agents intensions and actions. Finally, an example is given to illustrate the arithmetic of fusion and the propatation of new informaiton through the network, which shows the feasibility of using the network for solving problems of plan recognition.更多还原