基于排挤遗传算法的入侵检测方法  

Intrusion detection method based on crowding genetic algorithm

作　　者：刘文涛[1] 胡家宝[2] 

机构地区：[1]武汉工业学院计算机与信息工程系 [2]武汉理工大学计算机学院

出　　处：《计算机工程与应用》2010年第33期91-93,97,共4页Computer Engineering and Applications

摘　　要：传统遗传算法在入侵检测系统中构造的规则单一,导致检测率低,为了构造更加精确和完备的入侵规则,提出了一种基于确定性排挤遗传的规则构造算法,该算法使用确定性排挤来产生下一代种群,能够有效保持种群多样性,获得全部最优解。给出了算法的步骤和仿真,以网络数据集KDDCup99为对象,详细分析了利用该算法来生成入侵规则的具体实现过程,对染色体编码和适应度函数进行了设计和实现。最后通过实验证明了此算法的有效性,可以较好地获得入侵检测规则。The rules produced by the traditional genetic algorithm for intrusion detection system are too single and lead to the low detection rate.In order to get precise and complete intrusion detection rules,this paper puts forward a method based on deterministic crowding genetic algorithm.It can get the whole global optimal solutions and keep the diversity of the popu- lation.The basic steps and case simulation of this algorithm are presented.The concrete method of creating the intrusion detection rules using this algorithm is analyzed in detail using the KDDCup99 as the testing data set.The chromosome code and fitness function are proposed.The experiments show that the algorithm is efficient and can get better intrusion detection rules.

关 键 词：遗传算法 确定性排挤 入侵检测 KDDCup99 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]