基于贝叶斯网络的多agent分布式入侵检测技术  

Distributed Intrusion Detection Technology Based on Bayesian Networks & Multi-agent

作　　者：杜献峰[1] 

机构地区：[1]中原工学院计算机学院,河南郑州450007

出　　处：《华北水利水电学院学报》2007年第4期66-68,共3页North China Institute of Water Conservancy and Hydroelectric Power

基　　金：河南省教育厅资助项目(2006520020)

摘　　要：通过对分布入侵检测技术和agent技术的研究提出了一个基于多agent的分布式入侵检结构.结构中用Bayesian网络来表示对计算机网络进行攻击的知识和团树传播算法,将Bayesian网络划分为多个Bayesian子网,每个Bayesian子网转变为一棵团树;每个Bayesian子网有一组agent管理,agent通过Bayesian子网进行似然推理,从而能够实现在检测到入侵活动时提前发出警报.By studying the distributed intrusion detection and agent technology,an architecture is presented based on distributed intrusion detection and multi-agent. Bayesian Networks is used to present the knowledge of intruding computer networks and clique-tree propagation algorithm. Bayesian Network is divided into a few Bayesian sub-networks and one Bayesian sub-network is transformed a clique tree in agent reference algorithm. Every Bayesian sub-network is supervised by a set of agents and agent can do probability reference by Bayesian sub-networks. Intrusion alert is made when system eavesdrops intruding-activities.

关 键 词：MULTI—AGENT 贝叶斯网络 分布式入侵检测 

分 类 号：TP393.08[自动化与计算机技术—计算机应用技术]