Studies On Information Spreading And Intervention Strategies In Complex Networks

Real-world systems mostly are composed of a huge number of interacting individuals,and information spreading is one of the basic forms of interaction among these individuals.If we abstract these real systems into complex networks,then many spreading phenomena(for instance,the diffusion of rumors on social networks,the new technology promotion in business cooperation network,and the spreading of computer viruses on the computer networks)can be abstracted as information spreading on complex networks.It is of great practical significance to understand,model,and eventually intervene in the information spreading on complex networks.However,due to the complexity of network structure and information spreading mechanism,there are still many difficulties in the study of information spreading modeling and intervention.Firstly,in terms of modeling,the current modeling and theoretical analysis of information spreading on multiplex networks are yet to be improved.Secondly,in terms of intervention,current researches are mostly heuristic studies relying on central indicators,and people still have insufficient understanding of the specific intervention mechanism of intervention objects such as nodes or edges.To address the above issues,this dissertation will study the modeling and intervention of information spreading on complex networks.The specific contents and innovations are summarized as follows:(1)The modeling and theoretical analysis of information spreading on multiplex networks.In real-life,information can spread simultaneously on multiple platforms.Therefore,the research of information spreading on multiplex networks is of great importance;but there are still many theoretical difficulties in this field.On the one hand,the interlayer structure of multilayer networks is too complex to describe;On the other hand,the dynamic correlation between the neighbors in different layers is also difficult to describe.Current researches are weak in the theoretical analysis of these two aspects;thus,it is difficult to accurately predict the spreading threshold and outbreak size of the information.To conquer this problem,we first propose a novel information spreading model based on the multiplex network,namely the multiplex-network-based Susceptible-InformedRecovered model.This model emphasizes two points:(a)the multiplex network has complex inter-layer correlation;(b)the state of each individual will be affected by neighbors from different layers at the same time.On this basis,the theory of heterogeneous edge-based compartmental is introduced to comprehend the proposed spreading model,overcoming the problems of characterizing the inter-layer structure and describing the dynamic correlation between neighbors of different layers.With the help of this theoretical method,we can accurately predict the results and influencing factors of the information spreading in the multiplex networks.This part of the research has improved the theoretical analysis of the structural correlation and state correlation between the layers of the multiplex network,which can help to comprehend the information spreading process and its influencing factors in real-world multiplex networks,and lay a theoretical foundation for further research on information spreading intervention.(2)Node-based intervention method for misinformation spreading.Node is one of the basic components of complex networks.Controlling the nodes of the networks can help to contain the spread of misinformation.However,most of the current studies on intervention strategies for nodes are heuristic strategies based on the central indexes.It is difficult to analyze the intervention results and influencing factors of these heuristic strategies in the complex multiplex networks.Given this,we propose a reliable intervention strategy for the misinformation spreading on the multiplex networks,namely the node freezing strategy.Specifically,the node freezing strategy first assigns a weight value to each node in the network according to the degree of the node,and then freezing a fraction of nodes with high weight values so as to contain the spread of misinformation.Compared with the previous strategies,the node freezing strategy has two advantages:(a)the influence of the microstructure of the two sub-networks on the information spreading ability of the node is considered;(b)the intervention results and influencing factors of the node freezing strategy can be predicted and analyzed by the corresponding theoretical methods.This part of research can provide an effective decision basis for solving the problem of misinformation flooding in the real world,and also provide inspirations for the intervention research of other relevant spreading dynamics on multiplex networks.(3)Edge-based intervention method for information spreading.In addition to the node,the edge another basic element of complex network.Information in the complex networks is transmitted by the edges between nodes.Therefore,the edges in the networks play an important role in the process of information spreading.However,most of the existing studies on the intervention strategies for edges are heuristic strategies based on the central indexes.For instance,containing the information spreading by disconnecting the edges with the maximum degree product of the two ends of the nodes.These heuristic studies are not sufficient to help people understand the specific impact of different edges in the process of information spreading intervention.In view of this,we study the optimization problem of information spreading intervention for the single edge based on the traditional SIS information spreading model.The specific purpose of intervention includes the containing of misinformation spreading and the promotion of useful information spreading.In terms of containing the spreading of misinformation,we develop a perturbation method based on the discrete Markov chain of the SIS model.Using this theoretical method,we can calculate the decremental outbreak size after disconnected any edge in the network.In this way,we can find the optimal edge on the original network that can contain the spread of misinformation.In terms of promoting the spreading of useful information,we first propose an index to measure the impact of adding any potential edge by analyzing the evolution rule of the SIS model.Combined with the edge state equations of the SIS model,this proposed index can quantify the impact of adding any potential edge on the information spreading,so as to get the optimal potential edge that can facilitate the spreading of useful information.In this section,we solve the optimization problem of information spreading intervention for the single edge,which can help to understand the specific impact of adding or deleting any edge in the process of information spreading intervention and lay a theoretical foundation for further researches on information spreading intervention strategies based on multiple edges.