Application On Intrusion Detection Based On Multi-swarm Cooperative Evolution Particle Swarm Optimization Algorithm

With the popularity of the Internet and the rapid development of informationnetworks, the emergence of a large number of network security issues, previouslysimple firewall technology, has been unable to solve the intricate networksecurity issues. Network security mainly toward variability, intelligence andhidden development are the main problem. To enhance network security, manyorganizations do not leave a more robust and more aggressive strategies and plans,and intrusion detection is one of the effective methods to solve this problem, itcan provide real-time intrusion detection reports to the user while taking afavorable protective measures.With the network security field to intrusion detection technology depth study.Active defense and intelligent handling technology and its protection are a fewcharacteristics of distributed intrusion detection techniques for intrusion detectiontechnology has become a hot research field of network security. When the systemis subject to external attacks, intrusion detection technology can help to maintainthe safety and operability, and intrusion detection system can continue to providecritical services.This paper focuses on intrusion detection, classification and detectionmethods also introduces and describes the development process of particle swarmalgorithm in detail. This article focuses on the cooperative particle swarmalgorithm and a variety of group co-evolution particle swarm optimization stepsand processes, through the classic function, showing MSPSO algorithm canquickly find the global optimum fitness function, it has strong global searchability. It also can make up for the premature convergence into local minima.Using this algorithm KDD CUP99intrusion detection rules extracted data sets, focusing on KDD CUP99training data set and test data sets for testing.Algorithm rules KDD CUP99test data sets used in the training set and testset, the results show that the algorithm is applied to detect the training set toachieve an average rate of95.251%, the false positive rate is only4.862%average detection rate appling to the test set average of88.557%, an average of7.058%false positive rate. This algorithm can solve the detection rate, the highrate of false positives in certain extent and effectively detect network anomaliesin the data; the algorithm is effective and practical extraction rules.