摘要
为了避免军用软件在投入部队使用后频繁出现失效问题,进而造成训练及战时的损失,提出一种优化SVR算法的军用软件可靠性预测方法,旨在军用软件大量投入部队使用前,做好军用软件自身的可靠性预测工作。进行度量元采集,并对度量元进行数据预处理;通过引入相关性的概念,提取出与软件可靠性指标相关程度较大的度量元;再通过网格寻优法优化SVR中的参数;通过案例对比分析,证明了基于优化SVR算法的军用软件可靠性预测方法可以提高软件可靠性的预测精度。
In order to avoid military software being put into the army,the problem of frequently occurred failure during the use of military software,which resulted in the loss of training and wartime.We should do well the reliability prediction before military software is used in a large number of troops.Therefore,a military software reliability prediction method based on SVR algorithm is proposed.This paper firstly collects the metrics and preprocesses the metrics.Secondly,by introducing the concept of correlation,it extracts the metrics that have the higher degree of correlation to software reliability index.Thirdly,the parameters in SVR are optimized by grid search method.Finally,through case analysis,it is proved that the military software reliability prediction method based on the optimized SVR algorithm can indeed improve the prediction accuracy of software reliability.
作者
马振宇
张威
吴纬
韩坤
刘福胜
MA Zhen-yu;ZHANG Wei;WU Wei;HAN Kun;LIU Fu-sheng(Army Academy of Armored Forces,Beijing 100072,China;Unit 63963 of PLA,Beijing 100071,China)
出处
《火力与指挥控制》
CSCD
北大核心
2019年第11期151-155,共5页
Fire Control & Command Control
基金
军队科研计划基金资助项目(2015H04)
关键词
相关系数
度量元
可靠性指标
网格寻优法
支持向量回归
correlation coefficient
metric
reliability index
grid search method
support vector re gression
