A NOVEL PARTICLE SWARM OPTIMIZATION WITH MUTATION FOR INTRUSION DETECTION
This paper presents a novel system of applying particle swarm optimization with mutation (MPSO) for intrusion detection (MPSO-IDS). By applying mutation operator to the PSO algorithm near the end of run, the advanced MPSO algorithm can not only maintain the fast converging characteristics in the early phase, but also avoid the premature convergence phenomenon in the later phase. The MPSO algorithm is employed to derive a set of constraint-based detectors from training data, and the support-confidence framework is utilized as fitness function to judge the quality of each detector. The generated detectors are then used to detect abnormal IP packets by any-r intervals matching rule in the training data and testing data. Using the data sets of KDD CUP 1999, the experiment results show that the proposed MPSO-IDS can avoid the premature convergence problem effectively, and achieve better detecting rate than PSO-IDS and SGA- IDS.
intrusion detection particle swarm optimization mutation operator constraint-based detectors any-r intervals matching rule support-confidence framework
GUIHUA SUN YUFANG ZHANG ZHONGYANG XIONG
College of computer, Chongqing University, Chongqing 400044, China
国际会议
开封
英文
435-442
2006-10-15（万方平台首次上网日期，不代表论文的发表时间）