负指数分布排错时间的软件可靠性模型
A Software Reliability Model with Exponential Repair Time
-
摘要: 传统的软件可靠性模型大多假设排除错误的时间忽略不计,基于假设过于苛刻,影响了模型应用的范围和精度。因此在JM模型的基础上,提出了排错时间为负指数分布的软件可靠性模型及本模型的极大似然参数估计方法。通过用马尔可夫过程描述系统的状态,分析了模型的数值求解方法,并讨论了使用在本模型中的失效预测、剩余错误预测和等待排出的错误数预测公式。同时利用实际的工程数据对本模型的应用过程进行了验证。Abstract: Conventional software reliability models assume that the time to remove a fault is negligible. This rigid assumption limits the models' prediction precision and their application. Based on JM model, we propose a software reliability prediction model involving fault-remove time which followed exponential distribution. We provide the maximum likelihood parameter estimation equations of this model. We also provide several reliability prediction equations by using markov process. Illustrations of the proposed model by using real data are provided.