考虑级联失效的复杂负载网络节点重要度评估

Evaluating Node Importance Considering Cascading Failure in Complex Load-networks

首先提出了一个基于负载重分配的复杂负载网络级联失效模型,与现有模型的区别在于:(1)针对每个节点定义了一个“过载函数”;(2)用节点权值演化代替网络拓扑结构的演化.在级联失效模型基础上,提出了考虑级联失效的复杂负载网络节点重要度评估方法,给出了节点重要度的新定义.该方法有助于发现网络中一些潜在的“关键节点”.最后的实验分析验证了该方法的有效性和可行性.

This paper presented a model with ＂overload function＂ for cascading failure. The main differences with respect to previous models are as follows： （1） ＂overload function＂ is defined for each node, according to the value of ＂overload function＂, one node has three states： success, overload, failure. After the load decreases, an overloaded node can be success again. （2） The evolution of topology is replaced by the evolution of value of ＂overload function＂ during the process of cascading failure. It＇s needless to delete the failure nodes and its edges, the load will avoid the failure nodes automatically and the decrease of network performance will be reflected by network efficiency. An evaluation method of node importance considering cascading failure is proposed, and its algorithm is presented. A new definition of node importance is proposed. The most important node is the one whose failure results in the largest decrease of networks efficiency at the end of cascading. The evaluation method can help us to find some potential critical nodes which are sensitive to the efficiency of networks but not so important intuitively. Final example verifies its efficiency and feasibility.