Study On Complex Networks Node Pivot And Invulnerability Of Networks

With the continuous development of human society,increasingly complex systems have become ubiquitous.Most of the complex systems in the real world can be abstracted and analyzed as complex networks,from wireless communication systems to transportation systems,from ecosystems to logistics supply chain systems,by analyzing the important characteristics of complex networks can help us to better explore the characteristics and laws in these complex systems.Network nodes are an important part of complex network topology,and the importance analysis of complex network nodes,as one of the hot topics of network science research in recent years,has far-reaching research value.In order to explore the important hub nodes in complex networks more effectively,based on the traditional Leader Rank node importance ranking algorithm,the ranking process is improved by considering the influence of local clustering coefficients on node information propagation and the association properties of each node in the association network,taking into account both the nodes’ own properties from the microscopic and the structural characteristics of complex networks from the macroscopic perspective;in order to study the In order to study the influence of association structure on the network robustness after cascade failure in complex networks,a new cascade failure model of association networks is proposed,the influence of nodes in the network is mined by the Leader Rank algorithm considering local clustering coefficients based on associations,a new load redistribution strategy for failed nodes in association networks is formulated,and the robustness of heterogeneous networks is analyzed using theory and simulation.This paper focuses on the following main research topics and highlights its corresponding achievements:(1)In order to more effectively mine the hub nodes in key positions in complex networks,an improved Leader Rank node importance ranking algorithm(C-Leader Rank algorithm)based on the traditional Leader Rank node importance ranking algorithm is proposed by considering the influence of local clustering coefficients on the information dissemination of network nodes.Based on the real country port logistics data,an improved TOPSIS(Technique for Order Preference by Similarity to an Ideal Solution)comprehensive evaluation method and gravity model are used to establish a port-type complex logistics network.The experimental results show that the proposed improved algorithm can effectively explore the hub nodes that take into account the central and "bridging" attributes of the complex network nodes.(2)A large proportion of complex networks obtained from real system modeling have association properties.Considering the high clustering and heterogeneity of associations,based on the proposed algorithm in(1),an association-based Leader Rank algorithm(CCL algorithm)considering local clustering coefficients is proposed to identify the influential nodes in association networks.The importance analysis of nodes is performed by the CCL algorithm and other classical node ranking algorithms in Dolphin network and self-built cross-border airline network,and the proposed improved algorithm is further verified to be effective in identifying high-impact nodes in the association network using SI model.(3)Further to study the cascade failure problem in association networks and improve the overall network robustness,a new cascade failure model considering the association structure is proposed based on the node influence and similarity in association networks obtained by the CCL algorithm proposed in(2),and a robustness analysis is performed for the cascade failure process of association networks.Theoretical analysis and simulation results show that network heterogeneity has a great influence on the cascade failure process and node similarity has a positive influence on network robustness;after cascade failure occurs,the network has stronger robustness if the load quota is redistributed to same-community neighbors than cross-community neighbors.