基于改进模拟退火算法的软件可靠性模型参数估计方法  

Software Reliability Model Parameter Estimation Method Based on Modified Simulated Annealing Algorithm

作　　者：温晓玲 孙智孝[1] 姜梦岑 艾笑天 WEN Xiao-ling;SUN Zhi-xiao;JIANG Meng-cen;AI Xiao-tian(AVIC Shenyang Aircraft Design&Research Institute,Shenyang 110035,China)

机构地区：[1]航空工业沈阳飞机设计研究所,辽宁沈阳110035

出　　处：《测控技术》2022年第8期66-70,103,共6页Measurement & Control Technology

基　　金：航空工业联合基金“十三五”项目(6141B050301)。

摘　　要：为提升现有软件可靠性模型的拟合性能和求解精度,结合软件可靠性模型求解特征,提出一种改进的模拟退火算法。在此基础上,提出基于改进模拟退火算法的软件可靠性模型参数求解方法(简称为MSAE法),并将新方法应用于4组失效数据集。工程应用结果表明,与最大似然估计(MLE)法、和声搜索(HS)算法和蚁群(AC)算法相比,MSAE法可有效改善软件可靠性模型参数求解不收敛的情况,并且可以有效提升现有软件可靠性模型的拟合性能。In order to improve the fitting performance and solution accuracy of the existing software reliability model, combined with the solution characteristics of the software reliability model, an modified simulated annealing algorithm is proposed.On this basis, a software reliability model parameter solving method based on improved simulated annealing algorithm(MSAE) is proposed, and the new method is applied to four sets of failure data sets.The engineering application results show that compared with the maximum likelihood method(MLE),harmony search algorithm(HS) and ant colony algorithm(ACA),MSAE method can effectively improve the non-convergence of software reliability model parameters, and can effectively improve the fitting performance of existing software reliability models.

关 键 词：软件可靠性 软件可靠性增长模型 模型参数估计 模拟退火算法 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]