Analysis And Optimization For Monolayer And Multilayer Networks Robustness

Complex networks exist in every corner of people's life, therefore, scholars have never stopped the pace of exploration about complex networks. The structure of complex networks is carrier of interactive information and transmission resources; its robustness is directly related to the economics of the country and human life. So, how to construct robustness networks is one of the key branches in the field of complex networks in recent years. Most exising research is to ensure that the degree of each node does not change in process of optimizing the network structure. The social network structure is always changing in reality, and the degree of nodes in the network may be changing at any time. Meanwhile, the paths between nodes in the network are critical, they are the channels for any information's transmission. Moreover, different types of networks may interdependent, which is also known as the multi-layer network(Network of Network, NON), and the connectivity between them is an important part of interdependence of a network. First of all, the thesis study the method to optimize network robustness under different sides of attacks; and then, a method is proposed to optimize the natural connectivity of networks; finally, multi-layer network connection is explored again and gives quantitative analysis. The main work of this thesis is summarized as follows:1:During analyzing the robustness of network, set a new constraint rule, select three different link-attacks to attack the network. Moreover, we use four types of networks and a real network as initial networks, propose a sample heuristic method(SH) to optimize the networks under these different link-attacks. The experimental results show that the values of robustness measures for the optimized networks starting from different initial networks are similar under different link attacks, but the network topologies may be different. That is to say, networks with different topologies may have similar robustness in terms of the robustness measures. We also find that the optimized networks obtained by one link attack may not robust against other link attacks, sometimes, even weaker than the original networks.2:Adopted the natural connectivity as a measure of network robust, propose a method to optimize the natural connectivity of scale-free networks using simulated annealing, moreover, in the optimization process, keep the degree of each node invariant. In the simulated annealing, design an operation of exchanging edges, namely increase loop 3(IL-3). The performance of simulated annealing is validated on two real world networks and artificial networks, and experimental results show that the simulated annealing method clearly outperforms other algorithms.3:Considering mutual connection between real world complex networks, quantitative analyses on the robustness of NON that constituted by four different types of network are carried out. Useful suggestions to create NONs in the future are given. In the experiments, four different types of networks are used to consist different NONs. Nodes of these NONs are attacked randomly and maliciously. The experimental results show that NONs that are constituted by regular networks and small world networks have a great capability to endure the malicious attack, and the NONs that are constituted by scale-free networks and other networks have a great capability to endure the random attack.