摘要： 检测子生成是阴性选择算法(NSA)的关键步骤，现有检测子生成算法存在检测子生成效率不高、合格检测子冗余。针对该问题，改进检测子生成算法，对随机生成检测子进行二次耐受减少非自体空间的交叉覆盖区域，引入变异机制降低随机检测子与自体的碰撞概率。实验结果表明，改进算法对Probe攻击和DoS攻击的检测均优于NSA算法。

关键词: 检测子, 变异, 阴性选择, 入侵检测

Abstract: Detector generation is the key step of Negative Selection(NSA), but current detector generation algorithms exist redundancy detector and the lower efficiency, aiming at this problem, this paper improves detector generation algorithm, adopts the second tolerance in order to decrease the cross-cover area of non-self spaces for random generation detector, introduces the mechanism of mutation to decline the collision probability between randomly generated detector and self. Experimental results show that improved algorithm is superior to NSA algorithm for detecting Probe attack and DoS attack.

Key words: detector, mutation, negative selection, intrusion detection

中图分类号: 

• TP393.08