免疫检测模型中检测器的激励响应机制研究  

The Research of the Stimulation and Response Mechanism of Detector in Immune-Based Detection Model

作　　者：孙夫雄[1] 黄天戍[2] 

机构地区：[1]中南财经政法大学信息学院,武汉430074 [2]武汉大学电子信息学院,武汉430072

出　　处：《计算机学报》2006年第11期1929-1936,共8页Chinese Journal of Computers

基　　金：公安部科研项目基金(200342-823-01);中南财经政法大学科研启动金(00008673;XXXYKT2006003)资助.

摘　　要：文章研究与借鉴免疫系统中B细胞的生成、演化和检测的过程,针对当前IDS免疫模型研究中存在的不足,提出了检测器在不同检测阶段中的演变形式和行为方式,并给出了检测器的异常触发、活化概率函数以及相互间激励或抑制的算法与机制,同时针对活化响应发生链式反应、检测器的识别域之间交迭覆盖以及检测器规模等问题,建立了优先权和率先活化机制.实验结果验证了检测器激励响应机制一方面实现了检测器动态更新和联想记忆的功能,另一方面摒弃了盲目地随机产生检测器的传统方式,提高了免疫检测模型的自学习和自适应的能力.With the inspiration of the process of naive B-cells＇ generation, evolution and detec tion, the paper designs the evolving forms and action manners of detectors in different detection stages according to the limitation existed in current immune based IDSs. Meanwhile a mechanism of Abnormity Triggering and the probability functions for stimulation and restraint are given in detail. In order to solve some problems such as chain response, overlapping and detectors＇ size, a mechanism of Prior Order ＆ Prior Activation is built. The experimental results demonstrate the novel mechanism can not only realize the dynamic update and association memory of detectors, but also discard the traditional manner to produce detectors blindly and randomly, which improves the adaptability and self-learning of immune-baSed detection model.

关 键 词：入侵检测 自然免疫系统 激励响应 联想记忆 异常触发 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]