Agent-based Modeling And Experimental Analysis Of Social Networks Evolution

Social network and its evolution involve many aspects such as social life,engineering deployment,academic research and so on.The existing researches reveal some of their attributes,but it is not enough to describe groups in different backgrounds and different regions.Social networks are interlinked with Internet and Protein-Protein Interaction network.The research of social network evolution has reference value for explaining the operation mechanism of nature and other networks.Based on the innovative social network model and individual decision model,this thesis uses agentbased method to simulate the evolution process of network,and analyzes the influencing factors.The main work of this thesis is as follows:(1)Social network model.On the basis of exploring the basic characteristics and evolution characteristics of classical social networks,the concepts of similarity and interlock are introduced,the corresponding indexes are given,and a social network model based on Similarity-Interlock effect is proposed.(2)Decision-making model.Inspired from the existing decision-making theories of various disciplines,combined with the influence of individuals and groups on the decision-making subject,the basic framework of the decision-making system is constructed,and then the individual decision-making process is divided into multiple stages to simulate the individual decision-making process from the perspective of psychology.(3)Agentbased modeling experiment.On the basis of the proposed network model and decision model,the agent-based method is used to simulate the community growing process.The effectiveness and rationality of the proposed model are verified by experiments,and the factors affecting community growing are analyzed.The results show that the proposed social network model can better reflect the topological properties of the real network.The decision-making model is more accurate and efficient than the traditional model.In the agent-based modeling experiment,the selection of different initial nodes,changing the importance of the community and the structure of the network have different effects on the community growing.