一种改进的阴性选择免疫算法  

An Improved Intrusion Detection Negative Selection Immune Algorithm

作　　者：张海英[1] 管洪娜[1] 潘永湘[1] 

机构地区：[1]西安理工大学自动化与信息工程学院,陕西西安710048

出　　处：《西安理工大学学报》2005年第3期306-309,共4页Journal of Xi'an University of Technology

摘　　要：针对阴性选择免疫算法中字符串空间收敛效果差以及运行成本高的缺点,依据Forrest阴性选择算法提出了一种新的改进算法。该算法基于模式概念建立检测器库,检测器库存储了所有与自体不匹配的字符串个数,使得所占用的空间比所有与自体不匹配的字符串所占用的空间显著减少。采用了一种新的检测失败率算法公式,使得检测器的个数能够自适应自体规模的变化。理论分析与仿真结果表明,该算法与原算法相比,具有更快的运行速度和更低的检测失败率。Based on the study of Forrest negative selection algorithm, a new method is presented to improve the existing negative selection algorithm with bad convergent effect upon the space of string and high running costs. This algorithm is based on the detector library established in terms of the pattern concept. This detector library can store all the numbers of self-unmatched strings,thus making the occupied space be reduced more apparenthy than the space occupied by the self-unmatched strings. Accordingly,a kind of new algorithm formula to detect the failore rate is adopted so as to make the number of detectors be able to adapt to the selfbobu scale changes. The theoretical analysis and simulation results indicate that this new algorithm in comparison with the original algorithm ,is of rapid operation speed and lower detection failure rate.

关 键 词：模式 检测器库 自适应 

分 类 号：TN912.3[电子电信—通信与信息系统]