Research On Centrality Ranking Of Nodes In Multilayer Network And Mutual Propagation Of Multiple Viruses

The research of complex network mainly includes network structure evolution,network data processing,network dynamic process and network control.Among them,the dynamic process of complex network is very rich,such as various propagation processes,cascading failures,synchronization phenomena and so on.These network dynamic processes also lead to the related network control research,such as the spreading and immunization of computer viruses in the Internet,the spreading and control of infectious diseases in social networks,the spreading and control of rumors in social networks,the prevention and control of successive failures in power networks,and the spreading and prevention of crises in economic networks.These researches are related to the application of complex network theory in real systems,and they are the hot spots in the related fields of network science.This dissertation mainly studies the virus spreading behavior and immune strategy on the complex network,and analyzes the real system in real life with the multi-layer network as the carrier,and considers the key factors according to the internal spreading mechanism of multiple virus interactive spreading behavior,such as the location of nodes in the multi-layer network,the importance of each layer of network,the interaction of spreading sources,etc.In this dissertation,we make use of the interactive Markov and build a dynamic model of multi-layer network with multiple propagation sources.Combining with Monte Carlo simulation,we analyze the propagation dynamic characteristics of multi-layer network,such as propagation threshold,propagation scale,etc.Then we propose a new propagation immune strategy of multi-layer network,which aims to achieve better comprehensive immune effect in multi-layer network with lower global immune density.The main work and innovation achievements of this dissertation are as follows:1.This research is studying on the comprehensive central sorting algorithm of individuals with multiple attributes in multi-layer network.Based on the classic Google page ranking algorithm Page Rank,this dissertation studies the comprehensive central ranking algorithm for individuals with multiple attributes in multi-layer network.Using random walk model for reference,a new multi Page Rank algorithm based on multi-layer network is proposed.Monte Carlo simulation method is used to simulate the evolution process of multiple Page Rank in two-layer artificial network and two-layer real network.Theoretical and simulation results show that the importance distribution of nodes in multi-layer network is different from that in traditional network.Comprehensive consideration of multiple attributes of individuals is helpful to achieve the order of the comprehensive centrality of nodes in the real society.2.The research is studying on the algorithm of individual synthesis central sorting in the evolution process of multi-layer network considering the characteristics of shunting.Based on the possible overlapping characteristics of the connecting edges of each layer network in multi-layer network,a multi Page Rank algorithm of nodes considering the shunt is proposed based on multi-layer network.Through Monte Carlo simulation method,the evolution process of Page Rank in double-layer artificial network and double-layer real network is simulated respectively to verify the accuracy of theoretical derivation.The theoretical and simulation results show that the comprehensive centrality of nodes in multi-layer networks considering shunting is different from that in traditional networks.According to different situations,choosing more realistic multi Page Rank algorithm can identify the most influential nodes more accurately.3.The multi-layer network virus transmission dynamics is modelled and analyzed considering the interaction of multiple transmission sources.Based on the classic SIS model of virus transmission,considering the interaction of multiple viruses in different transmission ways,a new(SIS)~m model based on multi-layer network is established by using the micro Markov chain method.Five groups of virus cross propagation experiments are carried out by Monte Carlo simulation method.Each group simulates the process of cross propagation between two kinds of viruses in the double-layer network,and analyzes the critical value and final infection scale of each virus under the influence of cross propagation.Then we compare the simulations with the critical value and final infection scale of the corresponding virus independent propagation,and verify the difference between the interactive propagation model of multiple viruses in the multi-layer network and the independent propagation model of a single virus in the traditional network.Theoretical and simulation results show that the process and results of the interactive transmission of multiple viruses in multi-layer networks are related to the interactive transmission coefficient between viruses.4.The immune strategy of multi-layer network is analyzed and researched that based on the characteristics of specific network topology.First of all,based on the related theorem and characteristics of continuous time Markov chain,using the theory of dynamic average field,a multi-layer network virus interactive propagation model is established;then,the most influential nodes are identified by using the comprehensive central ranking algorithm of nodes in the multi-layer network;then,an immune strategy of multi-layer network virus interactive propagation is derived from the two In order to achieve the goal of global target immunity to a variety of interactive viruses in multi-layer network.In this way,we can get a better global immune effect against multiple viruses with a lower global immune density.Through Monte Carlo simulation,five groups of virus cross-propagation immune experiments were carried out.Each group simulates the cross-propagation process of immune two kinds of virus in the double-layer network.We analyze the immune critical value and infection scale changing curve of each virus when it was immunized with multiple targets,and compare with when it was immunized with multiple randomization,to verify the feasibility of the proposed multi-layer network multi-virus interactive transmission immune strategy.Theoretical and simulation results show that the proposed multi-layer network multi virus interactive transmission immune strategy can achieve a better effect of controlling multi virus transmission at the same time by immunizing fewer nodes.