基于MIDBO-SVR的网络安全态势评估方法  

Network security situation assessment method based on MIDBO⁃SVR

作　　者：陈秋琼[1] 徐华志 熊伟男 刘卫丽[1] CHEN Qiuqiong;XU Huazhi;XIONG Weinan;LIU Weili(Department of Information System,PLA Dalian Naval Academy,Dalian 116018,China)

机构地区：[1]海军大连舰艇学院信息系统系,辽宁大连116018

出　　处：《现代电子技术》2025年第5期101-106,共6页Modern Electronics Technique

基　　金：中国博士后科学基金项目(2015M572694,2016T90979)。

摘　　要：针对当前网络安全态势评估模型收敛速度慢、准确度不高的问题,提出一种多策略融合改进的蜣螂优化(MIDBO)支持向量回归机(SVR)的网络安全态势评估方法。利用混沌精英反向学习策略初始化蜣螂种群,提升种群多样性和质量;采用自适应权重的黄金正弦策略对滚球蜣螂位置进行更新,以平衡局部挖掘能力和全局搜索能力;引入Lévy飞行策略改进偷窃蜣螂位置更新公式,提高算法跳出局部最优能力;借鉴鲸鱼优化位置更新方法改进繁殖、觅食行为,提高算法的收敛速度和寻优精度。通过MIDBO优化SVR,获得最佳核函数参数和惩罚因子,构建MIDBO-SVR网络安全态势评估模型。实验结果表明,MIDBO-SVR评估模型与APSO-SVR、DBO-SVR和SVR评估模型相比,具有更高的分类准确率和运行效率,综合性能最优。The current network security situation assessment models have slow convergence speed and low accuracy,so a multi-strategy improved dung beetle optimization(MIDBO)algorithm is used to optimize the support vector regression(SVR)machine,and on the basis of the optimized SVR machine,a network security situation assessment model is established.The chaotic elite opposition-based learning(EOBL)strategy is used to initialize the dung beetle population,so as to enhance the diversity and quality of the population.A golden sine strategy with adaptive weight is adopted to update the position of the rolling dung beetle,so as to balance the local excavation ability and global search capability.The Lévy flight strategy is introduced to improve the position update formula of the thieving dung beetle,so as to enhance the algorithm′s ability to escape from local optima.The position update method of whale optimization is referenced to improve the reproduction and foraging behavior,so as to increase the convergence speed and improve optimization accuracy of the algorithm.The MIDBO is used to optimize SVR machine,so as to obtain the optimal kernel function parameters and penalty factors.On the basis of the above,a network security situation assessment model based on MIDBO-SVR is constructed.Experimental results show that the assessment model based on MIDBO-SVR has higher classification accuracy and operational efficiency in comparison with the assessment models based on APSO-SVR,DBO-SVR and SVR.To sum up,the proposed model has the optimal comprehensive performance.

关 键 词：网络安全 态势评估 多策略 DBO 支持向量回归机 自适应权重 

分 类 号：TN711-34[电子电信—电路与系统]