多值因果图的一种推理算法

A Reasoning Algorithm in Multiply Value Causality Diagram

该文针对多值因果图存在的两个困难:不严格满足概率论;将其用于实际问题时,推理结果可能出现错误。提出了一种基于因果影响可能性分配的推理算法。该算法对多值因果图进行了补充定义,使多值因果图能够兼容单值因果图;定义了事件变量状态可能性,及其精确计算方法和近似计算方法;给出了将多值因果图转化为单值因果图的方法和步骤,并推导了相应的计算公式;给出了收到证据后,感兴趣变量状态的后验概率计算方法。以核电站二回路系统中蒸汽发生器故障诊断因果图为例,展示了该算法推理计算的全过程。实例表明,该算法能够有效地克服多值因果图存在的困难,其推理过程严谨,计算结果符合实际情况。

The multiply value causality diagram has two difficulties,which does not satisfy with probability theory rigorous and the inference result maybe be error when applying it to experiment.Due to overcoming the difficulties,a reasoning algorithm based on possibility allocation is presented in this paper.The multiply value causality diagram is supplementally defined,so it can be compatible with a single value causality diagram.After defining the possibility of an event variable,this paper gives the precise computation method and approximate computation method.The computation step and method,which transform a multiply causality diagram into a single causality diagram,and the computation formula of linkage event are given.Finally,the method updating belief after accepting the evidence is given.A causality diagram about fault cause of a steam generator in the nuclear power plant has been established as an example.The example illustrated all the reasoning process.