摘要： 在基于人工免疫理论入侵检测系统中，变异算法的缺点导致了较高的误报率和漏报率。该文提出了一种变异机制，能够加快进化速度，保存具有优势特征的物种，提高检测效率和准确性。该机制定义了自体/非自体的概念和形式化描述，给出了成熟细胞动态方程、亲和力累积方程。实验结果表明，当k取40附近值时，只要参数合适，TP值就能稳定在95％以上。该模型具有良好的实时性、自适应性和准确性，为构建新一代高效、合理的网络安全系统提供了一种有效方案。

关键词: 网络安全, 人工免疫系统, 入侵检测, 变异算法

Abstract: Aiming at the deficiencies of current mutation algorithm based on artificial immune system, such as fault positive is high and the efficiency is very low, an improved evolution optimization with mutation algorithm is proposed. This new algorithm introduces evolution operator into the mutation rule during the cells mutation process to improve the detection efficiency and to overcome the shortcoming of the local optimum. And the concepts and formal definitions of self/non-self are given, mature-lymphocyte lifecycle, and affinity accumulation process are presented. Experimental results show that when the k set about 40 with the parameter adapted, the TP value can be stabilized on 95％. And the algorithm greatly enhances the response rate and precision of detection and the proposed model has the features of real-time processing, self-adaptively, thus providing a promising solution for intrusion detection.

Key words: network security, artificial immune system, intrusion detection, mutation algorithm

中图分类号: 

• TP301.6