The Study Of The Robustness Of Edge-based Interdependent Networks

Complex network is an important tool for studying real-world complex systems.With the continuous deepening of research and needs of practical purpose,the research on complex networks has gradually advanced from the isolated network to the interdependent network.The interdependent network is used to describe coupled complex systems which are increasingly common in modern society.For such a coupled system,the functioning of one of its subsystems depends on the functioning of other subsystems coupled to it.For example,the coupled system composed of the power grid network and the information network which act as the two subsystems of the whole system.However,the interdependent networks studied in the previous work mainly consider the dependence across subsystems through nodes in respective subnetworks,such as the dependence between the stations of power grid network and the server computers in the information network.In fact,many coupled systems in real life are coupled through edges connecting nodes rather than nodes themselves,which could be essentially different from those interdependent through nodes.For example,the interdependence between the economic chain and the trade chain of companies is based on the edge interdependency.In order to study the above-mentioned system based on edge interdependence,this paper introduces an edge-based interdependent networks(EIN)and conducts a systematic study on its robustness.Compared with the previous node-based interdependent network(NIN),the EIN considered in this paper is composed of multiple network layers,and the layers are connected through the interdependence of edges.Through the research of this paper,it is found that in average speaking,the nodes at the ends of an edge could reach more nodes than one node,which makes EIN more robust than NIN.This conclusion is proved by introducing a quenching network model with detailed theoretical analysis.This important conclusion reveals that the number of the nodes that a node at both ends of an edge can reach to is a key index for determining the robustness of EIN.Therefore,this thesis introduces a concept of the compound excessive degree(CED)of an edge to depict this key index,and further systematically study the impact of this index on the robustness of EIN under different conditions.Further research finds that the reason why the robustness of EIN is stronger than that of NIN in the case of random attacks essentially comes from the fact that the CED of the edge is larger than the degree of the node.On this basis,the thesis studies the effect of different attack methods based on CED on EIN.It is found that in the case of random attacks,the network based on the positive correlation interdependence of CED has the strongest robustness,while the network based on the negative correlation interdependence of CED has the weakest robustness.However,in the case of malicious attacks,the positive correlated EIN is more vulnerable than the negative correlated EIN when the edges with larger CED are removed first.This case is reveres when smaller CED edges are removed first.In summary,this paper studies the edge-based interdependent networks(EIN),and reveals the key factor that determined robustness of EIN,namely the compound excessive degree(CED).This thesis further systematically studies the effect of CED on the robustness of EIN in different situations.This thesis provides a new perspective on the study of interdependent networks,and its findings contribute to a deep understanding of the robustness of real-world coupled complex systems.