基于Copula函数模型的构成软件可靠性预测方法

Component Software Reliability Prediction Method Based on Copula Function Model

利用传统方法预测构件软件可靠性时,没有构建缺陷模型进行缺陷分析,存在构件软件可靠性预测时间延迟高、CPU开销大和平均路径覆盖率差的问题。于是提出基于Copula函数模型的构件软件可靠性预测方法,建立软件缺陷模型和故障树,对软件缺陷进行分析。基于二维焦元建立一个普通BPA证据变量,利用贝叶斯方法解决证据变量中样本量较少的问题,并从中取得Copula优质函数后构建联合BPA。通过构建的联合BPA和取得的最优Copula函数实现构件软件可靠性预测。实验结果表明,通过对比预测时间延迟、CPU开销和平均路径覆盖率,验证了所提构件软件可靠性预测方法的实用性和可行性较强。

During the software reliability prediction of components,the lack of defect analysis results in long prediction time delay,high CPU overhead and poor average path coverage.Therefore,the component software reliability prediction method based on the Copula function model has been reported.A software defect model and fault tree were established to analyze software defects.An ordinary BPA evidence variable was founded via two-dimensional focal elements.The Bayesian method was introduced to increase the number of samples in evidence variables in order to obtain copula quality function,and then construct joint BPA.BPA and the Copula function were used to predict the reliability of component software.The experimental results show that this method has a short prediction time delay,low CPU overhead and excellent average path coverage.