一种用于网络入侵检测的免疫克隆策略

An Immune Clonal Strategy for Network Intrusion Detection

网络入侵检测当前面临的主要问题是如何迅速有效地检测出未知模式的入侵。借鉴生物免疫系统中的自进化学习机制,我们设计一种免疫克隆算法,该算法以生物免疫的自我非我识别为基础,进一步引入免疫克隆学习机制以提高算法对入侵模式识别的效率和正确率。论述参数的设置,并且系统不再简单地丢弃穷举法中与Self匹配的候选检测器,而是对它们进行进化,引导它们偏离Self集合,生成检测器。论述免疫克隆算法的具体细节,并完成相应的验证实验。实验表明该算法具有较好的识别未知模式的能力。

The main problem of current network intrusion detection is how to recognize the intrusion of unknown patterns rapidly and efficiently. Inspired by the self evolution learning mechanism in the biological immune system, we designed art immune algorithm, which is based mainly on the self and non - self recognition. We also used the learning mechanism to improve its effi- ciency and preciseness for recognizing intrusion patterns. We describe the parameters of the algorithm and do not reject the detec- tors simply which match with the self, but instead with the evolutionary methods to generate the detectors. This paper introduced the details of the immune clonal algorithm and accomplished corresponding experiments. The experiment results given showed that this algorithm had a good ability to recognize unknown patterns.