计算机仿真2024,Vol.41Issue(8) :386-393.

基于Copula函数的故障相依软件系统可靠性模型

Failure-Dependent Software System Reliability Model Based on Copula Function

郑周桃 杨剑锋 贺孟兰
计算机仿真2024,Vol.41Issue(8) :386-393.

基于Copula函数的故障相依软件系统可靠性模型

Failure-Dependent Software System Reliability Model Based on Copula Function

郑周桃 1杨剑锋 2贺孟兰1
扫码查看

作者信息

  • 1. 贵州大学数学与统计学院,贵州 贵阳 550025
  • 2. 贵州大学数学与统计学院,贵州 贵阳 550025;贵州理工学院大数据学院,贵州 贵阳 550025
  • 折叠

摘要

传统的软件系统可靠性模型常常容易忽略子系统之间的故障相依性从而不能很好的反映对整个软件系统的可靠性建模分析.故根据实际情况考虑了屏蔽数据的存在且基于Copula函数考虑了子系统之间的失效相依性,进而提出了具有故障相依的软件系统可靠性叠加模型来解决此类相依性问题.又可靠性叠加模型的参数估计是难点,所提模型参数复杂,涉及到的似然函数变量维度高难以求解,故从可靠性出发,利用期望最小二乘算法(ELS)和最小二乘估计(LSE)来求解模型参数的近似估计值.最后通过一组实际数据和一组仿真数据进行数值验证,从最后的结果分析发现所提模型结果更加贴合于实际,且能很好的拟合累计故障数,从而验证了所提模型的有效性.

Abstract

The traditional software system reliability model often overlook the fault dependence between subsys-tems,and cannot reflect the reliability modeling and analysis of the entire software system well.Therefore,according to the actual situation,the existence of shielded data and the failure dependence between subsystems are considered based on Copula function,and then a software system reliability superposition model with fault dependence is proposed to solve this kind of dependency problem.In addition,parameter estimation of reliability superposition model is diffi-cult.The proposed model parameters are complex and the likelihood function variables involved too high dimension to solve.Therefore,from the perspective of reliability,expected least squares algorithm(ELS)and least squares estima-tion(LS)are used to solve the approximate estimates of model parameters.Finally,a set of actual data and a set of simulation data are used for numerical verification.From the analysis of the final results,it is found that the results of the proposed model are more consistent with the reality,and can well fit the cumulative fault number,thus verifying the effectiveness of the proposed model.

关键词

可靠性/故障相依/屏蔽数据/期望最小二乘算法/仿真数据

Key words

Reliability/Failure-dependent/Masked data/Expected least squares algorithm(ELS)/Simulation data

引用本文复制引用

基金项目

国家自然科学基金资助项目(71901078)

贵州省电力大数据重点实验室(黔科合计Z字[2015]4001)

出版年

2024
计算机仿真
中国航天科工集团公司第十七研究所

计算机仿真

CSTPCD
影响因子:0.518
ISSN:1006-9348
段落导航相关论文