基于D-S理论的因果图区间分析方法

Interval Analysis Method of Causality Diagram Based on D-S Theory

【目的】针对实际应用中事件发生概率具有不确定性的特点,研究解决获取事件发生概率精确值的问题。【方法】运用D-S理论提出了因果图区间分析方法,避免了获取概率的精确值,使用似然函数和信任函数分别作为因果图中基本事件区间概率的上限和下限。针对出现异常现象的事件先确定出可能导致其发生的基本事件,再分别求出这些基本事件的后验概率,然后利用相对优势度对区间数比较大小。【结果】区间数最大的基本事件就是故障源。【结论】用区间数来表示事件发生的概率值,解决了事件发生概率精确值的不确定性问题。

[Purposes]According to the actual application of event occurrence probability has the characteristics of uncertainty,In order to solve the problem of obtaining the accurate values of event occurrence probability.[Methods]The D-S theory is used to put forward the method of causality diagram of interval analysis.Applying this method avoids obtaining the accurate values of difficulty of the event occurrence probability,the likelihood function and the belief function in the D-S theory are used as upper and lower boundary on the interval probability of fault occurred in basic events in causality diagram analysis.When abnormal phenomenon is occurred,firstly,we should determine the basic events that may lead to its occurrence,and then calculate the posterior probability of basic events.[Findings]Interval numbers are compared based on the relative superiority degree.Basic event of the most interval number is fault source.[Conclusions]Using interval number to denote the probability of event value,solves the problem of uncertainty event probability.