Modeling Of Bounded Confidence Opinion Dynamics With Spatial Constraints

Public opinion as a kind of complicated social phenomenon in social system hasattracted a great deal of interest from researchers in different scientific fields. Theevolutionary model of public opinion is becoming a research hotspot recently in the studyof public opinion. A large number of models were proposed based on simplified dynamichuman complex behaviors, but failed to describe the influence of physical distancebetween two individuals on the evolution of public opinion with respect to complexity.In this thesis, we introduce spatial constraints into modeling the evolution process ofpublic opinion. Due to the finite resources of individuals, physical distance is introducedto form the local communication of individuals. Combining with the researches inpsychology and sociology, new factors including individual attributes, personalitycharacteristics and public authority are expressed quantitatively. By using the agent-basedmethod, an evolutionary model of public opinion based on physical distance is developedfor the reality of individual behavior.Simulation experiments are carried out on our developed platform for research onevolutionary model of public opinion. Assuming that confidence radius and physicaldistance radius are both identical, our numerical results prove the effectiveness of ourmodel by inheriting the three dynamic characteristics, i.e. fragmentation, polarization andconsensus. The convergence time for opinion evolution is greatly increased. Physicaldistance radius hinders the consensus of public opinion in some extent but promote theformation of fragmentation. When physical distance radius is beyond a critical value it hasnothing to do with the opinion dynamics, only adjusting the convergence time and opinionnoises. Particularly, when confidence radius is in a specified interval, the result of thedynamics is consensus under a small physical distance radius. And then assuming thatconfidence radius and physical distance radius are both randomly chosen according topower law distribution, we find that the heterogeneity of confidence radius and physical distance radius can promote the opinions to achieve consensus. Through comparisons withtwo classical bounded confidence models, our results are in good agreement with the newfeatures in the expansion of human communication, and provide a comprehensiveexplanation for the evolution of public opinion in real social system.