摘要
针对现有的粒子群优化(PSO)算法大多存在早熟收敛、容易陷入局部最优值的问题,提出了一种新的协同粒子群优化(CPSO)算法。该算法拥有两个子群,一个用于全局搜索始终保持粒子多样性,另一个用于局部搜索保证搜索精度,通过相互协同合作在全局最优值附近实现精确搜索。最后把该算法应用到动态聚类入侵检测,通过优化聚类半径和聚类阈值,对训练数据进行正、异常类聚类,然后用测试数据进行攻击检测。试验结果表明该算法较粒子群和突变粒子群(MPSO)算法性能明显提高。
Aiming at the problems of premature convergence and easy to fall into local optimum value of existing particle swarm optimization（PSO） algorithms,a new collaborative particle swarm optimization（CPSO） algorithm is proposed.CPSO algorithm has two subgroups,one subgroup is used for global search always keep particle diversity,the other one is used for local search guarantee search precision.So precise search is realized nearly the global optimal value by mutual cooperation.Finally,the proposed algorithm applied to intrusion detection based on dynamic cluster.Through the optimization of clustering radius and clustering threshold,the training data is classified as normal and abnormal clustering.Then test data is used to attack detection.The results show that CPSO algorithm has a marked improvement in performance over the traditional PSO algorithm and improved mutation particle swarm（MPSO） algorithm.
出处
《燕山大学学报》
CAS
2013年第2期124-128,147,共6页
Journal of Yanshan University
基金
河北省教育厅基金资助项目(2007493)
关键词
粒子群优化
协同粒子群
动态聚类
入侵检测
PSO
premature convergence
collaborative particle swarm
dynamic cluster
intrusion detection
