基于不确定故障树的软件可靠性分析方法

Software Reliability Analysis Method Based on Uncertain Fault Tree

故障树是用来评估系统可靠性或风险的有效方法之一,其中基本事件故障发生概率通常为常数或随机变量.在实际软件可靠性测试过程中,可得到的故障数据量非常少且存在大量的认知不确定性,因此使用传统故障树评估方法是不合理的.为此,本文提出基于不确定故障树的软件可靠性分析方法.故障树中基本事件发生故障的信度用不确定测度表示,以不确定环境下“与门”和“或门”的运算法则计算系统的可靠度.最后,对某装备的软件系统进行可靠性分析及灵敏度分析,使该方法的合理性得以验证.

The fault tree is one of the effective methods used to evaluate the reliability or risk of the system,in which the probability of occurrence of each event failure is usually a constant or random variable.In the actual software reliability testing process,the amount of available fault data is very small and there is a large amount of epistemic uncertainty.Therefore,it is unreasonable to use the traditional fault tree evaluation method.For this reason,in this article we propose a software reliability analysis method based on uncertain fault trees.The belief degree of basic events failure in the fault tree is first expressed by uncertainty measure,and the reliability of the system is then calculated by the algorithm of“AND”and“OR”gates in an uncertain environment.Finally,the reliability analysis and sensitivity analysis of the software system of a certain equipment are conducted to verify the rationality of the method.