人工免疫系统中人工抗原提呈细胞的设计  

Design of Artificial Antigen Presenting Cells in Artificial Immune Systems

作　　者：杨鹤[1,2] 郑幸[3] 梁意文[2] 

机构地区：[1]湖北第二师范学院计算机学院,武汉430205 [2]武汉大学计算机学院,武汉430072 [3]武汉军械士官学校,武汉430074

出　　处：《小型微型计算机系统》2012年第5期1121-1125,共5页Journal of Chinese Computer Systems

基　　金：国防基础项目(A1420080183)资助;国家自然科学基金项目(61170306)资助;湖北省教育厅科研项目(B20113010)资助

摘　　要：反向选择模型是人工免疫系统中的经典模型,但存在训练代价高、覆盖率低的问题.危险理论模型是人工免疫系统中新引进的模型,在降低训练代价,提高覆盖率上具有优势.建立人工免疫危险理论模型必须解决的关键问题之一是:在保持人工免疫系统自适应特征的前提下,设计人工抗原提呈细胞.本文从抗原提呈细胞的生物学原理出发,设计了人工免疫系统中的人工抗原提呈细胞,给出单个人工抗原提呈细胞的结构、细胞上Toll样受体的结构,以及人工抗原提呈细胞群体的结构;设计了人工抗原提呈细胞群体的生命周期.初步的实验结果证明所设计的人工抗原提呈细胞具有自适应发现恶意软件的能力.Negative Selection is the classical model in Artificial Immune Systems.High training costs and low coverage are the most serious problems of this model.Danger Theory is a new model in Artificial Immune Systems.It has advantages in reducing training costs and improving coverage.One of the key problems which must be solved in building Danger Theory model is designing Artificial Antigen Presenting Cells,at the same time the adaptability of the model should be maintained.This paper presented a set of design of Artificial Antigen Presenting Cells,including the structure of individual Antigen Presenting Cell,the structure of Toll-Like Receptors,the structure and life cycle of Antigen Presenting Cells population.Preliminary experiments showed that the design had the ability to find malicious software adaptively.

关 键 词：人工免疫系统 危险理论 人工抗原提呈细胞 恶意软件检测 

分 类 号：TP309[自动化与计算机技术—计算机系统结构]