改进遗传算法优化神经网络的入侵检测研究  

Research on Intrusion Detection of Optimized Neural Network Based on Improved Genetic Algorithm

作　　者：杨云峰[1] 唐凤仙[1] YANG Yunfeng TANG Fengxian(School of Computer and Information Engineering, Hechi University, Yizhou, Guangxi 546300,Chin)

机构地区：[1]河池学院计算机与信息工程学院,广西宜州546300

出　　处：《河池学院学报》2017年第2期77-83,共7页Journal of Hechi University

基　　金：2016年广西高校中青年教师基础能力提升项目(KY2016YB380);河池学院智能计算与模式识别重点实验室科研项目

摘　　要：针对大数据时期的互联网安全日趋严重的形势,构建一个利用粗糙集和改进新种群生成方式的遗传算法优化BP网络的网络攻击检测模型,通过粗糙集进行除噪降维,改进的遗传算法的新种群生成方式是通过选择部分最优父代个体进行交叉变异后替换父代最差个体产生新种群,BP网络获得改进新种群生成方式的遗传算法提供更佳初始参数,BP网络的检测速率慢及局部最小的问题获得更好解决。本模型经过仿真测试证明其增长检测正确率,减少检测时间。Based on the situation that internet security is becoming increasingly serious and important in the age of big data, a network attack detection model has been built which optimizes the BP network by using the rough set and improved genetic algorithm of new population generation mode. The new population generation mode produced by the improved genetic algorism is the mode through whicha new population is generated by selection of the optimal parent individuals undertaking crossover mutation to replace the worst parent individuals. The improved genetic algorism of new population generation mode provides better initial parameters to BP network and supplies a better solution to the slow detection rate and local minimization problem of BP network. Simulation tests of this model have proved the improvement of detection accuracy and reduction of detection time.

关 键 词：粗糙集 改进遗传算法 BP神经网络 全局最优 

分 类 号：TP37[自动化与计算机技术—计算机系统结构]