智慧城市环境下加权网络级联故障与社团划分  

Cascading failures of weighted network and community detection in smart city

作　　者：李志清[1] 

机构地区：[1]广州行政学院市情研究所,广东广州510070

出　　处：《计算机工程与设计》2017年第6期1479-1484,共6页Computer Engineering and Design

基　　金：广东省自然科学基金项目(07001802)

摘　　要：针对智慧城市中各种系统、网络中不同重要性连边导致的级联故障问题,提出一种改进的面向社团划分的加权网络级联故障模型。将节点初始负载定义为考虑边权重占比的邻居节点介数之和,节点容量计算附加社团因素。数值模拟加权网络的级联故障发现,边介数加权网络的社团特性较弱,等权重加权网络的社团特性较强,网络的社团特性越明显,网络抵制故障能力越强;对节点附加二次容忍值可以显著提升网络抵制故障能力;桥梁节点比非桥梁节点具有更大的级联故障破坏力。To solve the problem that different edges of smart city have different importance because of complexity of systems and networks,which is more suitable for weighted network depicting the strength of node pairs.An improved cascading failure model of weighted networks considering lacks of existing schemes was presented,in which two innovations were proposed,the initial load was defined as the sum of neighbors edge weight ratio and the community factor of node was added to the capacity.Results of numerical simulation of different weighting networks show that,weighted network using betweenness centrality for edges has weaker community features,while weighted network using equivalence for edges has stronger community features.The more obvious characteristics of the community of network have,the stronger resistance capacity of the network has.Additional secondary tolerance value can significantly enhance network capacity to resist failure.Bridge node has greater destructive power of cascading failure than the non-bridge node.

关 键 词：智慧城市 复杂网络 社团划分 级联故障 加权网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]