摘要： 将生物免疫机制引入入侵检测系统，设计了一个基于免疫代理的入侵检测系统。系统探测和响应采用层次结构，各Agent既相互独立又相互协作，游走于各网络节点间，检测分布式的攻击。该文介绍了免疫算法的寻优机理，抗体扩增和抑制、记忆单元更新、亲和度和浓度计算等关键技术，确保了抗体的多样性，改善了未成熟性的收敛。

关键词: 免疫代理, 进化, 入侵检测系统

Abstract: Inspired by the biological immune mechanism, this paper designs an intrusion detection model based on immune agents, which uses the level construction of sensor and response scheme. The immunity-based agents are independent and collaborate each other and roames among network nodes to detect the distributed intrusions. It introduces the artificial immunization algorithm to seek the superior mechanism, the antibodies promotion and suppression technology, the update of memory, the affinity and the density. The experiment shows that this system guarantes the antibodies multiplicity, improves premature convergence.

Key words: Immune agent, Evolution, Intrusion detection system