级联网络的软件可靠性预测

研究软件可靠性准确预测问题,软件存在动态失效性,且引起软件运行失效的原因具有随机性,不同可靠性模型预测相同软件得到的结果不一致,通用性比较差,导致预测精度低。为了提高软件可靠性预测精度,提出一种级联网络的软件可靠性预测模型。采用4种经典软件可靠性模型的输出作为BP神经网络模型的输入,利用各种单一预测模型的优点,建立一种新的级联软件可靠性模型。仿真结果表明,级联网络模型具有更高的预测精度和通用性,验证了级联网络预测模型对软件可靠性预测的有效性和良好的应用前景。

Prediction of Software Reliability Based on Cascaded Neural Network

Research software reliability prediction problem.Softwares have the feature of dynamic failure,different reliability models have different prediction results,and the prediction accuracy is low.In order to improve the prediction accuracy of software reliability,this paper put forward a cascade network software reliability prediction model.Using four classic software reliability models as the output of BP neural network models,using the advantages of each forecast model,a new cascade software reliability model was set up.Simulation results show that cascade network model has higher precision of prediction and generalization,and the cascade network prediction model of software reliability prediction has good application prospect.