Multi-Agent Intrusion Detection System Based on Immune Principle
Speed-bottleneck and bandwidth problems in traditional distributed intrusion detection system are solved by multi-agent, but false negative rate and false positive rate are still high. In order to solve the above problems, immune principle and multi-agent are combined, then immune agent is constructed and intrusion detection model based on immune agent is builded. An improved dynamic clonal selection algorithm is proposed. Operating principle of the system and realization of each agent are illustrated in detail. The proposed model and algorithm are simulated by KDD’99 datasets. Key parameters are optimized. Compared to other results, the proposed method has low false positive rate and higher detection rate in Dos, Probing, U2R, and R2L attacks.