入侵可能路径预先估算方法研究与仿真

发布时间：2018-03-21 05:34

本文选题：入侵　切入点：路径　出处：《计算机仿真》2014年08期 　论文类型：期刊论文

【摘要】：在网络入侵检测中,对入侵可能路径进行预估,都是以后发式入侵预防为主,无法形成先验预判,主要因为入侵特征存在较大伪装性,没有办法提取固定的入侵路径预判特征。传统的入侵检测方法都是基于固定入侵特征形成入侵判断的依据的,以后期发现为主,对网络的破坏已经形成,造成入侵检测缺陷较大。提出采用粒子群特征多层解析算法的入侵可能路径预先估算方法。根据主成分分析相关理论,获取对应的特征权重矩阵,得到入侵特征对角元素,完成入侵前期特征提取。根据粒子群特征多层次解析相关理论,获取初始粒子群,针对粒子群中不同层次的所有粒子进行交叉和变异运算,搜索入侵可能性最大的路径,实现入侵可能路径预先估算。实验结果表明,利用改进算法进行入侵可能路径预先估算,能够极大地提高估算的准确性和实时性,满足网络安全的实际需求。
[Abstract]:In the network intrusion detection, the prediction of the possible path of the intrusion is mainly based on the later heuristic intrusion prevention, which can not form a prior judgment, mainly because of the larger camouflage of the intrusion characteristics. Traditional intrusion detection methods are based on fixed intrusion characteristics to form the basis of intrusion judgment. In this paper, a method of estimating the possible path of intrusion based on particle swarm optimization (PSO) multilayer analytic algorithm is proposed. According to the principal component analysis (PCA) theory, the corresponding feature weight matrix is obtained and the diagonal elements of intrusion feature are obtained. According to the correlation theory of multi-level analysis of particle swarm characteristics, the initial particle swarm optimization is obtained, and the crossover and mutation operation of all particles in different levels of particle swarm is carried out to search the path with the greatest probability of intrusion. The experimental results show that the accuracy and real-time of the estimation can be greatly improved by using the improved algorithm to estimate the possible path of intrusion, and the actual demand of network security can be satisfied.
【作者单位】：北京中医药大学计算机教研室;
【分类号】：TP393.08

【相似文献】