多Agent的创新网络入侵检测方法仿真研究  

Simulation research of innovation network intrusion detection method based on multi Agent

作　　者：傅荣[1] 佘朝晖[1] 

机构地区：[1]湖南大学工商管理学院,湖南长沙410082

出　　处：《现代电子技术》2016年第24期39-43,共5页Modern Electronics Technique

基　　金：湖南省自然科学基金:面向互联网群体协作的知识网络演化研究(湘基金委字(2012)4号)

摘　　要：在多Agent的创新网络中,容易出现周期性漏洞和链路反馈溢出,为网络病毒的入侵提供了时隙。通过多Agent的创新网络入侵进行有效检测,可以实现对病毒的拦截和识别,提高网络安全性能。提出一种基于局部特征尺度的多Agent下创新网络入侵检测方法,分析网络结构和网络入侵节点分布特征,构建多Agent的创新网络入侵信号模型,进行入侵信号的滤波抗干扰处理,采用局部特征尺度匹配方法进行网络入侵的匹配滤波检测,实现网络入侵检测算法的改进。仿真实验结果表明,采用该方法进行网络入侵检测,准确检测概率高于传统方法,并具有较好的幅值响应性能,有效抑制了多Agent的创新网络中的旁瓣干扰,在网络安全防御中具有较好的应用价值。In the innovation network of multiple Agent, it is easy to occur periodic vulnerabilities and link feedback over- flow, which will provide a time slot for the network virus invasion. With effective detection of multiple Agent innovation network intrusion, the virus may be blocked and identified, and the network security performance can be improved. The innovation net- work intrusion detection method is put forward under multiple Agent based on local feature. The network structure and network intrusion node distribution characteristics are analyzed. An innovative network intrusion signal model of multiple Agent is built for the anti-interference filtering precessing of invading signal. The local feature matching method is used to conduct the matched filtering detection of network intrusion, and implement the improvement of network intrusion detection algorithm. The simulation experiment results show thai the network intrusion detection method has higher accurate detection probability than the traditional method, possesses good amplitude response performance, and can effectively restrain the sidelobe interference in innovation net- work of multiple Agent. It has a good application value in the network security defense.

关 键 词：AGENT 创新网络 入侵检测 网络安全 

分 类 号：TN926.34[电子电信—通信与信息系统] TP393[电子电信—信息与通信工程]