Research On Key Node Identification And Invulnerability Of Complex Networks Based On Cascading Failure

One of the most noticeable dynamic behaviors on complex networks is the cascading failure process,and the study of it is an important branch of the research on complex networks.The study of network cascading failure behavior is essential for realizing effective control of network catastrophic failures,preventing network crashes and ensuring the normal operation of the network,and helping to develop tools for predicting,preventing such failures and restoring normal network functions from it.Cascading failures are often caused by the failure of key nodes,and the robustness analysis of complex networks rarely considers the role played by important nodes,which has certain limitations.In the key node identification method,there are still many problems to be improved.In addition,because it is very expensive to improve the survivability of the network by changing the network topology,how to optimize the load redistribution strategy to control the network flow has become a research hotspot.In response to these problems,the main tasks are as follows:First of all,in the key node identification method,the traditional method is only for a specific network structure,which is one-sided and cannot be universally applicable,and the dynamic characteristics of the network are rarely considered in the identification process.To solving this problem,an optimization algorithm is proposed to identify key network nodes,and the network cascade failure process is introduced considering the dynamics of the network.Based on this,the network robustness measure is constructed to measure the network performance.This is the objective function,and the best point is used.The artificial fish school algorithm improved by the collection,chemotaxis and Levi’s flight strategy is used to optimize the search.Experimental simulation shows that the key node identification method in this paper is more accurate.Secondly,in the research of network survivability,this paper improve the cascading failure model,consider the real-time load changes of the network,and comprehensively consider the remaining capacity of the node and its neighbor nodes to define the real-time node status to improve the network load redistribution strategy,and find out where Under this new strategy,the network has more robust model parameters,and on this basis,combined with the key node identification method to study the network invulnerability effect.Experimental simulation shows that the improved cascading failure model can make the network more survivable in the face of conventional attacks,but the survivability is weaker in the face of the failure of key nodes.It is necessary to combine key node identification methods for key node protection.Finally,the effectiveness of the proposed key node identification method and the improved cascade failure model is verified in the actual network.The results show that the key nodes are protected before the cascade failure process occurs,and the improved load redistribution strategy which is implemented during the cascade failure process,can make the network have higher comprehensive survivability under different failure modes.