Research And Application Of Key Node Identification Theory And Method In Complex Networks

With the rise of complex network research,many research fields related to complex networks have been paid more and more attention by researchers.Among them,the identification of key nodes in complex networks is a research direction that has been gradually being valued.The research of nodes in networks can further understand the whole network structure and network functions.For different types of networks,the methods of studying key nodes are different.In order to design key node identification algorithms for different types of networks,this paper proposes four key node identification algorithms for different types of networks,considering the differences between nodes and links.At the same time,the applicability and accuracy of the algorithm are verified in different types of networks,and the algorithm is extended to practical network applications.The specific work of this paper is as follows:Firstly,through combing the research background and significance of the paper,the research questions are put forward,and the network is studied with the help of graph theory knowledge in complex network.For four different types of networks,unweighted-undirected network,weighted-undirected network,unweighted-directed network and weighted-directed network,an algorithm based on node adjacency information entropy is proposed to identify important nodes in different networks.The corresponding algorithm is designed according to different characteristics of different types of networks,in which weighted network considers the importance of node-to-node boundary weights.In the directed network,the different influence of the node's in-degree and out-degree is considered.Secondly,in order to verify the applicability and accuracy of the algorithm,this paper applies the algorithm to a simple small network,unweighted-undirected ARPA network and weighted-directed ARPA network,and verifies the applicability and accuracy of the algorithm by the node removal method.At the same time,compared with the related algorithms in different literatures,it is found that the number of subgraph networks obtained by this algorithm is large,and the scale of the network is small.The experimental results show the applicability and accuracy of this method in different networks.Moreover,in order to further verify and apply the algorithm in this paper,taking N city transportation and freight network as an example,the relevant parameters of N city transportation and freight network are obtained by using Pajek simulation software,and the adjacency information entropy of each node in the network is calculated by using Matlab,then the importance of different nodes in the network is obtained.At the same time,the practical significance of the important nodes in the network is analyzed by combining the geographical location map of N city transportation freight network and the functions of each node in the network.Finally,in order to understand the robustness and vulnerability of the network more comprehensively,the variation of network robustness under random faults and deliberate attacks is studied in N city transportation freight network.The robustness of the network is measured by three indexes: average path length,network clustering coefficient and network connectivity ratio.Path length and clustering coefficient can well reflect the impact of the two different ways on the network,while the network connectivity ratio shows a general effect.