PLC控制系统软件可靠性的神经网络预测  

Neural network based dynamic prediction of software reliability in PLC control system

作　　者：张家海[1] 徐耀群[2] 路明[1] 

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》2006年第4期662-664,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PLC控制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 PLC control software system. Based on software test record, the application of the neural network as a tool for dynamicly predicting the number of faults in program is explored. It indicates that the neural network can produce models with better fitting quality and predictive quality in comparison with traditional G - O model.

关 键 词：神经网络 PLC 软件可靠性 静态预测 

分 类 号：TP311.5[自动化与计算机技术—计算机软件与理论]