考虑故障检测与修正的软件可靠性增长模型

Software reliability growth model when fault detection and imperfect debugging progress considered

针对现有NHPP类软件可靠性增长模型基本假设中故障检测和不完美修正过程考虑不完善的现状,提出一种非齐次泊松过程软件可靠性增长模型.该模型既考虑故障本身特性,故障检测率是随时间递减的函数,又考虑了排错过程引入新错误的概率,故障总数函数是随时间变化的;同时,还考虑了不完全排错的情况,引入了一种与故障检测率具有相同变化趋势的故障排除率函数.利用一组标准化的失效数据集进行仿真验证,实验结果表明该模型具有更好的拟合效果和预测能力.

The fault detection and imperfect delay-debugging progress was considered imperfectly in the non-homogeneous poisson process （NHPP） software reliability growth model. Then, a NHPP software reliability growth model was brought forward. The own characteristic of fault was considered in the model and the fault detection rate （FDR） was a function decreasing with time. At the same time, the rate of bringing new error during the troubleshooting process was considered and failure function was the total number of changes over time. Meanwhile, the case of no fully troubleshooting was considered and a troubleshooting rate function which has the same trend with the fault detection rate was brought forward. The Effectiveness of the model was proved using a group of public effectiveness data. The result of the experiment shows that the model had better fitting results and predic- tion ability.