Cascading Failure And Centralization Research Of Scale-free Network

The study of scale-free network in my paper can be substantially divided into three closely related and deeply penetrate sides:quantified the statistical properties of the network through empirical; building the corresponding network model to explain the nature of these statistics properties; predicted the failure-resistant capability of the network based on the known topology and the rules of building the network. In the real networks, there are a variety of complex networks, such as transportation networks, power grids, communication networks, the networks of interpersonal relationship and so on. Optimized the classic scale-free network model in the real network to enhance the ability of resistance to the cascading failure, the work has great practical significance to improve the robustness of the network and its failure-resistant capability. Therefore, my paper focuses on the following three questions:First of all, the classic scale-free network model selects nodes with a certain probability in the global scope to priority connection, but it’s difficult to do this in the reality network. To solve this problem, in the second chapter of this paper builds a new scale-free network model based on the coupling coefficient through increasing two parameters of coupling coefficient and the attractive factor on the classic BA scale-free network model, and calculated the degree distribution of the evolution model in theory, the analysis found that it has obviously significant scale-free network characteristics. The simulation result also shows that the degree distribution obeys the power-law distribution and was more stability and broadly applicability.Secondly, chapter III researches the cascading failure problem caused by the random attacks and intentional attack on the scale-free network based on the coupling coefficient. Based on the Node Dynamic Model, follow the principle of that the load on the collapse nodes must be re-distributing locally and preferred, compared the global cascading failure phenomenon under the two attack strategies of the extended scale-free network relative to the typical model of scale-free networks, and explored the effectiveness of the two different attack strategies. Numerical simulation indicating that when adjust the parameter between a certain value in this model, attacking the smallest nodes in the network more easily lead to the cascading failure than attacking the largest node in the network, the extended network has been more robust.Studies have shown that there are great number of network phenomenon which are associated with the centralization degree of the entire network and the central nodes in the scale-free networks, such as the vulnerability of the Internet, the inequality between nodes in the aviation networks, the congestion of the transport networks, as well as how to reduce the dependence on the hub nodes of the protein interaction network and resist the spread of the virus in the network of human cells. Finally, the paper has researches the topology changes before and after the attacking to network model by several indicators of centralization. The case study shows that, the topology of the extended model before and after the attack does not change obviously, so indicating that it has a higher resistance to failure.