考虑S型测试工作量函数与不完美排错的软件可靠性模型  

A software reliability growth model considering an S-shaped testing effort function under imperfect debugging

作　　者：李海峰[1] 李秋英[1] 陆民燕[1] 

机构地区：[1]北京航空航天大学可靠性与系统工程学院,北京100191

出　　处：《哈尔滨工程大学学报》2011年第11期1460-1467,共8页Journal of Harbin Engineering University

基　　金："十一五"总装备部预研资助项目(513190702)

摘　　要：为准确描述测试工作量与不完美排错这2种重要测试过程因素对软件可靠性建模的影响,提升软件可靠性模型的拟合与预计精度,提出一种变形S型测试工作量函数,描述测试工作量增长速率随测试时间先增后减的S型增长趋势.在此基础上,分别提出考虑S型测试工作量函数与2种不完美排错假设的指数型非齐次泊松过程(NHPP)软件可靠性模型(即IS-TEFID1和IS-TEFID2).在2组真实失效数据集上,将新提出的2个模型与若干已存在的软件可靠性模型进行对比.实例验证结果显示,本文IS-TEFID2模型在2组失效数据集上的拟合与预计结果均显著好于其他模型,表明在建模过程中同时考虑变形S型测试工作量函数以及不完美排错可以有效地提升模型的拟合与预计性能.Software testing effort and imperfect debugging are two important testing factors which can effectively improve the fitting and prediction power of a software reliability growth model（SRGM）.Therefore,for describing the influence of these two factors on reliability in the modeling process more accurately,an S-shaped testing effort function（TEF） was proposed,which is also known as an inflected S-shaped TEF（IS-TEF）.This function is suitable and flexible for describing the S-shaped varying trend of the testing effort＇s increasing rate.Then two new NHPP SRGMs（IS-TEFID1 and IS-TEFID2） were presented by incorporating IS-TEF and two forms of imperfect debugging into the exponential-type NHPP SRGM,respectively.Finally,on two real failure data-sets,a case study was proposed for comparing the two proposed models with several representative SRGMs in terms of fitting and prediction power.The experimental results indicate that compared with these SRGMs,the proposed IS-TEFID2 yields the best fitting and prediction results for each data-set.That is,incorporating IS-TEF and imperfect debugging into the NHPP SRGM can provide accurate fitting and prediction results.

关 键 词：测试工作量 不完美排错 变形S型 软件可靠性增长模型 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]