Research Of Network Intrusion Detection System Based On Artificial Immune

With the rapid development of computer network, people have to face increasingly severe problem of network security. As an active and effective network security tool, intrusion detection can effectively make up the shortcomings of firewalls.Intrusion detection based on artificial immune system has referenced the complex information processing mechanism of immune system,and it become a research focus in the field of intrusion detection. Based on the study of theoretical knowledge that related to intrusion detection and artificial immune system, we research dynamic clonal selection algorithm for network intrusion detection.In the dynamic clonal selection algorithm, the remaining antigens that have been detected by memory detectors and mature detectors are sent to self set for immature detectors to tolerance.But the coverage of effective detectors is limited, not all the remaining antigens are self styles,some may be unknown type of intrusion. In order to reduce system holes, in this paper,we propose to use cluster on the remaining antigens,then analyse the small cluster to find hidden intrusion,and then put the unknown intrusion to memory detector set.It can not only purify self set,reduce system holes,but also enhance the learning ability of the system.In addition, it is unreasonable to delete redundant memory detectors by traditional random strategy. In order to improve spatial coverage of detectors,we use immune network algorithm to delete memory detector which has low stimulatation.What's more,it's inefficient to generate immature detectors randomly.To improve the quality of the production of immature detectors, in this paper,we use gene library strategy , mutate the deleted memory to generate immature detectors.Finally, we establish a network intrusion detection model based on the improved dynamic clonal selection algorithm. The simulation experiments results show that the improved algorithm can enhance the system's ability to discover unknown intrusion, improve the detection rate of intrusion detection.