基于网络效率最优的关键节点识别方法  

Key node recognition method based on optimal network efficiency

作　　者：冯小伟 胡聪 许畅 吴斌 倪平波 刘向军[1] FENG Xiao-wei;HU Cong;XU Chang;WU Bin;NI Ping-bo;LIU Xiang-jun(School of Electrical and Electronic Engineering,North China Electric Power University,Beijing 102206,China;Information and Communication Branch,State Grid Anhui Electric Power Company,Hefei 230000,Chin;Aostar Information Technologies Limited Company,Chengdu 610000,China)

机构地区：[1]华北电力大学电气与电子工程学院,北京102206 [2]国网安徽省电力公司信息通信分公司,安徽合肥230000 [3]四川中电启明星信息技术有限公司,四川成都610000

出　　处：《计算机工程与设计》2019年第2期328-335,共8页Computer Engineering and Design

基　　金：国家电网公司总部科技基金项目(SGTYHT/15-JS-191)

摘　　要：准确合理地找出网络中的关键节点并加以保护,对提高网络的鲁棒性、稳定性有着重要的意义。因此,从优化的角度提出一种基于网络效率最优的关键节点识别方法。通过网络结构赋权构造一种网络鲁棒性测度函数,以此为目标函数利用萤火虫算法进行优化搜索,其中在离散化的基础上,用佳点集的思想构造初始序列,采用分区寻优的方法快速搜索满足测度函数的最优序列,在搜索过程中通过增加可变全局吸引力和自适应随机项,使得优化搜索能够准确收敛,实现对网络关键节点的识别。实验分析结果表明,所提方法识别效果更佳,对不同的网络结构具有一定通用性,所用改进萤火虫算法收敛速度更快,准确性更高。Finding the key nodes in a network and taking measures to protect them is significant to improve the robustness and stability of the network.Therefore,a key node recognition method based on optimal network efficiency was developed from the perspective of optimization.A network robustness objective function was constructed by weighting the network,and firefly algorithm was utilized so that the optimal value was searched with the object function.The initial sequence was constructed according to good points set.Moreover,the optimal sequence that satisfied the objective function quickly was found with the help of partition optimization.During the search process,by adding variable global attraction and adaptive terms,the optimal search converged more accurately,so as to recognize the key nodes in the network.Results of simulations show that the proposed method performs better than the previous ones and it is applicable to different network structures.Furthermore,the modified firefly algorithm converges faster and more accurately.

关 键 词：网络效率 结构赋权 鲁棒性 佳点集 萤火虫算法 

分 类 号：TP308.02[自动化与计算机技术—计算机系统结构]