基于ADMPDE-WNN的网络安全态势预测方法  

A Method for Network Security Situation Prediction Based on ADMPDE-WNN

作　　者：黄仁全[1,2] 李为民[1] 张庆波[1] 董雯[3] 

机构地区：[1]空军工程大学防空反导学院 [2]中国人民解放军93942部队 [3]宝鸡石油机械有限责任公司

出　　处：《电光与控制》2013年第5期10-14,62,共6页Electronics Optics & Control

基　　金：国家科技重点实验室基金项目(9140XXXXXX110)

摘　　要：精确预测网络安全态势,对分析网络安全态势和趋势、调整安全策略具有重要意义。针对复杂网络系统规模变化,建立了层次化网络安全态势评估模型;在分析小波神经网络(WNN)基本原理及优缺点的基础上,将具有寻优能力强、收敛速度快的自适应局部增强微分进化算法(ADMPDE)与WNN算法相结合,提出了基于ADMPDE-WNN的态势预测方法。通过仿真实验分析,ADMPDE-WNN算法大大减小了训练误差和预测误差,提高了网络安全态势预测精度。基于ADMPDE-WNN的网络安全态势预测是一种科学、合理的预测方法,具有一定的理论意义和应用价值。The precise predicting of network situation is of great significance to the adjustment of network security measures. A hierarchical security evaluation model was proposed. The basic principle, advantages and disadvantages of the Wavelet Neural Network （WNN） were analyzed. Then, the Modified Differential Evolution Algorithm with Adaptive and Local Enhanced Operator （ADMPDE）, which had strong search ability and high convergence speed, was applied to the network security situation prediction together with WNN, and thus the ADMPDE-WNN was created. According to the simulation, both the training error and prediction mean error were decreased, and the prediction precision was improved. The method of the network security situation prediction based on ADMPDE-WNN was scientific and reasonable, and it is meaningful both in theory and application.

关 键 词：网络安全 态势预测 态势感知 微分进化算法 小波神经网络 

分 类 号：V271.4[航空宇航科学与技术—飞行器设计] TP824[自动化与计算机技术—检测技术与自动化装置]