Study On Consensus Of Collective Behavior From The Perspective Of Dynamics

With the continuous enrichment of natural science research methods,the evolution process of human opinion and its universal law have become the focus of new research in recent years.According to the process of the formation of different forms opinion,we can use statistical physical methods or complex network means to reveal its regular formation mechanism and detailed evolution process.This thesis mainly analyzes the electoral behaviors with distinctive characteristics by constructing a probabilistic evolution model.We also investigate the universal law of the evolution of the opinion by assuming different networks of interpersonal relationships.The specific work is as follows:Vote behavior of group mainly depends on understandings to an event and living environment of con-stituent,and also affected by the ideological orientation of the candidates.In this model,probability density evolution equation for describing the state of the constituents is suggested.Individual' s state is defined by attitudes(positive or negative),election intention(supporting candidate A or B or be a neutral)and the environ-ment(the corresponding probability density of different intentions).In certain parameters,election results are not depended on the initial distribution,but these can be affected by different combination of parameters,such as generalized coordinates of representing potential(x1,x2andx3)and transition coefficients(?-,12and ?+,32).There is a critical parameter ?-C,12 at which the constituent holding positive attitude is balanced with that of the neutral,and another critical parameter ?+,32C at which he constituents holding positive attitude is balanced with that holding negative attitude.The critical points are related to the corresponding coordinates,respectively in power and exponential manners.Therefore,from the perspective of candidate,it is effective way to strengthen the constituents' election intention when they publicized own political propositions by themselves.With the development of information technology,the way of human communication is more convenient.The complex interpersonal network leads to the diversity of opinion evolution,and the individual opinion is often influenced by the group where it belongs.In this thesis,based on the Hegselmann-Krause model,we consider influencing degree among interacting individuals,and introduce a convergence function to investigate dynamics mechanism of opinion consensus with different networks.Simulating results show that the opinion consensus can be affected by the network structures.The structure randomness is stronger,the time of opinion consensus is shorter.The increase of the convergence strength coefficient(?)is not conducive to the consensus of the group opinion,which leads to the shortening of the time when the group opinion evolves to the steady state.