基于Logistic增长神经网络模型的软件测试方法  

Software testing method for Logistic growth model based on neural network

作　　者：魏霖静[1] 宁璐璐[2] 练智超[3] 王联国[1] 侯振兴[4] 

机构地区：[1]甘肃农业大学信息科学技术学院,甘肃兰州730070 [2]南洋理工大学生物科学学院,新加坡639798 [3]南京理工大学计算机科学与工程学院,江苏南京210094 [4]南京大学信息管理学院,江苏南京210093

出　　处：《哈尔滨工程大学学报》2017年第4期646-651,共6页Journal of Harbin Engineering University

基　　金：国家自然科学基金项目(61063028;31560378);江苏省自然科学基金青年基金项目(BK20150784);中国博士后面上项目(2015M581800);甘肃省科技支撑计划(1604WKCA011);陇原青年创新创业人才项目(2016-47);2016年度甘肃省高校重大软科学(战略)研究项目(2016F-10)

摘　　要：软件可靠性评估性能直接影响软件测试的工作量,本文针对软件测试工作中的故障检测和校正处理问题,提出一种基于Logistic增长神经网络的软件测试方法。该方法考虑到软件工程的多样性,利用Logistic增长曲线构建神经网络模型完成故障检测,并结合指数分布校正时间完成故障校正过程。通过两组真实失效数据集(Ohba与Wood)的试验,将所提方法与现有的软件可靠性增长模型(software reliability growth model,SRGM)进行了比较。结果显示Logistic增长神经网络模型的模型拟合效果最优,表现出了更好的软件可靠性评估性能及模型适应性。Evaluation of software reliability performance directly affects the course of software testing.In this paper,we investigate fault detection in software testing to improve its performance.A software testing methods based on neural network of the logistic growth was proposed.Considering the diversity of software engineering,the proposed method using the logistic growth curve to construct the neural network model in order to complete fault detection.The proposed method combines the exponential distribution correction time in order to complete the fault correction process.Through the test of two sets of real failure data sets(Ohba and Wood),the proposed method is compared with the existing software reliability growth model(software reliability growth model,SRGM).The results confirm that the model fitting effect of the proposed logistic growth neural network model is optimal,demonstrating the adaption and better performance of the software reliability assessment model.

关 键 词：软件测试 可靠性评估 神经网络 软件可靠性增长模型 Logistic曲线 Wood数据集 Ohba数据集 故障检测 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]