基于神经网络算法的网络安全评价模型  

Network security evaluation model based on neural network algorithm

作　　者：周珑 郭威 王建永[1] 黄杰韬 ZHOU Long;GUO Wei;WANG Jian-yong;HUANG Jie-tao(Guangdong Power Grid Co.Ltd.,China Southern Power Grid,Guangzhou 510630,China)

机构地区：[1]中国南方电网广东电网有限责任公司,广州510630

出　　处：《沈阳工业大学学报》2018年第4期426-430,共5页Journal of Shenyang University of Technology

基　　金：国家自然科学基金资助项目(61033004)

摘　　要：为了准确地评价复杂网络的信息安全水平,设计了一个具有网络安全评价功能的GABP算法评价模型.该模型通过引入德尔菲法等方法,建立了科学的评价指标体系.基于这一体系,通过修改BP算法和遗传算法的若干步骤,选取适当的目标函数,并对算法的流程进行精确控制,最终得到了一个具备这两种算法优点的GABP评价算法.该评价算法的仿真结果表明,只需输入复杂网络的信息安全监测数据,算法就可以准确评价出该网络的信息安全水平.In order to accurately evaluate the information security level of complex network,a GABP algorithm evaluation model with network security evaluation function was designed. Through introducing the Delphi method and other methods,a scientific evaluation index system was established in the proposed model. Based on the established system and through modifying some steps of BP algorithm and genetic algorithm,the appropriate objective function was selected,the algorithm flowwas accurately controlled,and finally the GABP evaluation algorithm with the advantages of two algorithms was obtained. The simulation results showthat the proposed algorithm can accurately evaluate the information security level of network only through entering the information security monitoring data of complex network.

关 键 词：神经网络算法 遗传算法 德尔菲法 信息安全 安全评价 MATLAB软件 目标函数 适应度 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]