考虑测试工作量与覆盖率的软件可靠性模型

Software Reliability Model Considering both Testing Effort and Testing Coverage

为了进一步提升现有非齐次泊松过程类软件可靠性增长模型的拟合与预计精度,首先,提出一个同时考虑测试工作量与测试覆盖率的NHPP类软件可靠性建模框架.在此基础上,将变形S型测试工作量函数(IS-TEF)以及Logistic测试覆盖率函数(LO-TCF)带入该建模框架,建立了一个新的软件可靠性增长模型,即IS-LO-SRGM.同时,还对利用该框架进行建模过程中的两个重要问题进行了描述与分析,即如何确定具体的TEF和TCF以及模型参数估计.然后,在两组真实的失效数据集上,利用该建模框架建立了最为合适的增长模型,即IS-LO-SRGM,并将该模型与8种经典NHPP模型进行对比.实例验证结果表明,所提出的IS-LO-SRGM模型具有最为优秀的拟合与预计性能,从而证明新建模框架的有效性和实用性.最后,对不完美排错情况进行了初步的讨论与建模分析.

To further improve the fitting and prediction performance of the non-homogeneous Poisson process （MHPP） software reliability growth models （SRGMs）, this paper, as the extension work of the NHPP software reliability modeling framework which considers the TEF, will discuss how to integrate both TEF and TCF into the traditional NHPP software reliability modeling process. This is done in order to capture the integrated effect of testing effort and testing coverage on reliability estimation. First, a comprehensive modeling framework for incorporating the TEF and TCF together into the NHPP SRGMs is proposed. Recur to this framework, a new NHPP SRGM （named IS-LO-SRGM） with both the IS-TEF and logistic TCF （LO-TCF） is proposed. Meanwhile, two issues of this proposed framework are discussed respectively （i.e. how to select the most appropriate TEF and TCF for modeling and the parameter estimation）. Then, two case studies on two real failure data-sets are presented. The experimental results show that the IS-LO-SRGM nearly yields the best fitting and prediction results compared with the other comparison NHPP SRGMs for two data-sets. Thus, the applicability and effectiveness of this modeling framework are validated. Finally, the imperfect debugging phenomenon is also considered in the modeling framework for a further discussion.