基于粗粒度遗传算法的网络入侵检测系统  

Network Intrusion Detection System Based on Coarse-grained Model Genetic Algorithm

作　　者：李甦[1] 罗安坤[1] 

机构地区：[1]云南大学信息学院,昆明650091

出　　处：《计算机工程》2008年第13期166-168,171,共4页Computer Engineering

基　　金：云南省教育厅应用基础研究基金资助重点项目(03Z180A)

摘　　要：分析遗传算法在入侵检测系统中的可应用情况,提出一种基于粗粒度模型遗传算法的网络入侵检测系统。通过对协议特征的分析,找出有可能被非法利用和更改的特征属性,经过组合和编码后构成系统的初始种群,在各个处理器(终端点)并行地进行遗传算法的操作,使种群的进化在所有检测点同时进行,通过迁移相互交流,合理地设计适应度函数,使遗传"基因"的取舍和利用更加合理。实验数据表明,系统的检测率达到90%以上。This paper puts forward a Network Intrusion Detection System（NIDS） based on coarse-grained model genetic algorithm, after analysing the application of genetic algorithm in intrusion detection system. By analysing the protocol＇s property, finds out some characteristics which are often lawlessly changed and used, forms the original chromosome by assembling and coding, makes all processor（terminal points）carry genetic process by attributed manner, makes the evolution of chromosomes prosecute in all detection points simultaneity, and all processor can in by transplanting too. At the same time, redesign a reasonable fitness function, to let the use of ＂gene＂ be more reasonable. Experimental data shows the system＇s detection ability is above 90%.

关 键 词：网络入侵检测系统 遗传算法 种群 适应度函数 粗粒度模型 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]