Polymorphic Agents-based The Analysis On Evolutionary Mechanism Of Group Behavior In Social Networks

Group behavior is a widespread existence in social life.It is a social movement composed of multiple(three or more)individuals with common goals and mutual interaction.In a real society,positive group behaviors contribute to the development of society,and negative group behaviors hinder the development of society.Therefore,studying the evolutionary mechanism of group behavior can promote the sustainable development of society and have a important value.In recent years,with the complexity of group behavior increases,the traditional research methods based on mathematical models are difficult to make correct arguments for group behavior.However,with the rapid development of computers,Agent-Based Modeling methods are used to study the evolution of group behaviors.The mechanism begin to attract the attention of scholars.Through the simulation of microscopic individuals,it has played an important role in revealing the intrinsic mechanism of group behavior.In this thesis,through the analysis of typical group behavior application scenarios,three technical requirements for studying group behavior are extracted: modeling of social network structure,individual polymorphism in population and state transition model,and individual decision model in evolution of group behavior.In the social network structure modeling,by analyzing the attributes and link attributes of the nodes,a daisy-skull model of the social network is proposed;through the definition of individual polymorphisms and the analysis of state transitions,the process of evolution of group behaviors is constructed.Individual decision model.Finally,experimental analysis and argumentation of the daisy web model proposed in this paper were carried out on the experimental platform based on Netlogo,and the influencing factors of group behavior were analyzed.Experimental results show that the social network model constructed in this paper can better fit the topological properties of real networks.In the analysis of the influencing factors of group behavior,the individual's degree of paranoia,the number of neighbor nodes,the frequency of information release,and the network structure can well reproduce individual behavior decisions in the behavior of real groups.