Studies On The Opinion Dynamics Based On Adaptive Scale-free Networks

Complex networks are widely used in various systems of society. they have close relationship with human lives and affect each other. Public opinion is the sum of public ideas and views. The spread of public opinion is related to sociology, physics and complexity science and is a typical complex issue. Most complex networks show scale-free behavior in the real world. During the process of opinion spreading, two individuals with different opinions only influence each other when they are trustworthy. If they don’t trust each other, the individual can look for a trustworthy one to exchange their opinions. So, it has an important significance for the control and guidance of public opinion to study the public opinion dynamics on adaptive scale-free network, and to analyze the statistical properties of opinion evolution through computer simulation.This paper briefly summarizes the research overview of complex network,and then describes typimal models of public opinion dynamic, such as S(Sznajd) model, Voter model,MR(MajorityRule)model,D(Deffuant et al) model and the KH(Krause-Hegsekmann) model. Former three models are discrete binary ones, the latter two are bounded confidence ones with continuous opinion values. On this basis, we build a adaptive BA scale-free networks Deffuant et al public opinion dynamics model, in which convergence parameter is the function of degree values k, realize the numerical simulation and do the analysis. We studied the interaction between public opinion dynamics and network topology, including how opinion evolution affects degree distribution, statistical character of the opinion evolution with change of the confidence parameter ε, how network size N and confidence parameter ε affects system relaxation time, the analysis of public community.Due to the interaction between public opinion evolution and network topology, the initial inhomogeneous BA scale-free network becomes a uniform network; opinion evolution shows phase transition phenomena from plurality state to polarization state and then to the consensus state with ε increasing; system relaxation time and opinion community also have a close relationship to ε, showing a certain regular regularity. Further, we find the smaller the heterogeneity of the network, the greater confidence between individuals, the more conducive to the opinion evolution.