基于免疫危险理论的智能检测器及相关算法  

Intelligent detector and related algorithms based on immune danger theory

作　　者：谭敏生[1] 白曜华 周欢[1] 刘尖学 

机构地区：[1]南华大学计算机科学与技术学院,湖南衡阳421001

出　　处：《华中科技大学学报（自然科学版）》2015年第5期17-22,共6页Journal of Huazhong University of Science and Technology(Natural Science Edition)

基　　金：国家自然科学基金资助项目(61300234);湖南省教育厅科学研究重点资助项目(14A121);湖南省教育厅科学研究优秀青年资助项目(12B110);南华大学重点学科建设资助项目(计算机科学与技术);南华大学研究生科研创新资助项目(2013XCX19)

摘　　要：针对传统检测器检测率较低、误报率较高、自体库过于庞大的问题,设计了基于免疫危险理论的KRID(Katherine智能检测器),包括实时检测模块、记忆模块和自适应模块.提出了MSMA(二进制分段近似匹配方法)及KRNA检测算法,KRNA检测算法根据不同类型的危险信号使用不同的分段近似匹配方法进行分层检测.实验结果表明:KRNA检测算法具有较高的检测率、较低的误报率和冗余度;为进一步提高KRID的自适应能力,给出了DSAA(动态自适应算法),该算法获取各类异常信号的出现次数,将其与该类异常信号的危险频率求和取平均值后得到新的危险频率,根据新的危险频率调整检测细胞数量.实验结果表明:DSAA提升了检测器的自适应能力,减少了响应时间.KRID（Katherine intelligence detector）was designed according to the immune danger theory principle to solve the problem that traditional detector had low detection rate and high false rate,and its self set was huge and complicated.The MSMA（multistage similar miracle algorithm）matching method and KRNA（Katherine navigate algorithm）based on immune danger theory were also proposed.Different matching method was used according to different type of danger signal for layered detection.Experimental results show that KRNA has low detection rate,high false rate and reduced redundancy.The DSAA（dynamic self-adaptive algorithm）was proposed to solve the detector′low self-adaption ability problem,which got frequency of every kind of danger signals and danger frequency by taking the average of frequency sum of the danger signal plus primary frequency,then adjusted the detect cell number according to the new danger frequency.Experimental results show that DSAA can enhance the detector′s self-adaption ability and shorten the detector′s respond time.

关 键 词：智能检测器 免疫危险理论 物联网网络层 Katherine智能检测器 动态自适应算法 二进制分段近似匹配方法 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]