基于软件圈复杂度相关的不完美排错可靠性增长模型  

Imperfect Error Correction Reliability Growth Model Based on Software Circle Complexity Correlation

作　　者：张峰[1] 费琪[2] 李耀敏 ZHANG Feng;FEI Qi;LI Yaomin(No.91404 Troops of PLA,Qinhuangdao 066000;Jiangsu Automation Research Institute,Lianyungang 222061;No.722 Research&Development Institutey CSIC,Wuhan 430205)

机构地区：[1]91404部队,秦皇岛066000 [2]江苏自动化研究所,连云港222061 [3]中船重工集团第722研究所,武汉430205

出　　处：《舰船电子工程》2019年第11期135-141,共7页Ship Electronic Engineering

基　　金：国家自然科学基金项目（编号：61773384）;国防重点研发计划（编号：JSZL2017207B013）资助

摘　　要：首次将软件可靠性模型与软件圈复杂度进行了关联,将软件圈复杂度划分为三个等级:普通、复杂、特别复杂。针对不同等级的圈复杂度,分别从代码状况、可测性、可维护性、对测试人员能力的要求、对开发人员能力的要求5维角度进行了充分分析,从而给出了基于软件圈复杂度相关的不完美排错可靠性增长模型。在进行可靠性模型建模的同时综合考虑了故障的检测率、故障排错率、排错时故障的引入率三大因素,最后将该方法应用于一组实验数据进行分析,证明了该方法的可用性,同时将该模型的拟合数据与现有模型的拟合数据进行了比较,证明了该方法的有效性。The software reliability model is firstly correlated with the software circle complexity,and the software circle com plexity is divided into three levels:ordinary,complex and especially complex.According to different levels of the software circle com plexity,the software circle complexity is divided into five levels:code status,testability,maintainability,requirement for tester's abil ity and developer's ability.Five-dimensional analysis is required to give a reliability growth model of imperfect debugging based on the complexity of software circle.At the same time,the reliability model is modeled by considering three factors:fault detection rate,fault removal rate and fault introduction rate.Finally,the method is applied to a group of experimental data to verify the availability of the method.At the same time,the fitting data of the model and the fitting data of the existing model are used.The comparison shows that the method is effective.

关 键 词：软件圈复杂度 LOGISTIC增长模型 不完美排错 故障检测率 排错率 

分 类 号：TN911-34[电子电信—通信与信息系统]