组合导航软件可靠性的神经网络动态预测  

Neural network based dynamic prediction of software reliability in integrated navigation system

作　　者：张家海[1] 徐耀群[2] 

机构地区：[1]三江学院电气工程系,江苏南京210012 [2]哈尔滨商业大学系统工程研究所,黑龙江哈尔滨150028

出　　处：《哈尔滨工业大学学报》2005年第12期1691-1693,共3页Journal of Harbin Institute of Technology

基　　金：黑龙江省自然科学基金资助项目(F0215)

摘　　要：在组合导航系统大规模软件开发中,为了控制软件的质量,必须对软件模块的缺陷数进行预测.本文根据软件模块测试报告,用神经网络动态预测了软件模块的缺陷数.结果表明:神经网络预测软件缺陷数比传统的G-O模型有更好的质量拟合度和估测能力.Accurately predicting the number of faults in program modules is a major problem in quality control of a large scale integrated navigation software system. Based on software test records, the use of neural networks as a tool for dynamically predicting the number of faults in a program is explored. It indicates that the neural networks did produce models with better quality of fit and predictive quality than the traditional G -O model.

关 键 词：神经网络 组合导航系统 软件可靠性 动态预测 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]