| With the rapid development of artificial intelligence,internet,big data and mobile communication,technology has brought a series of new challenges to the regulation in man-machine integration network where the opinion dynamics research in social network is one of the most important directions.Experts from economics and management have discovered that the human opinion interaction behind social communication and management activities contains complex game relationships.Therefore,by exploring the behavioral rule of interactions between different individuals,we find that two complex social processes of opinion dynamics and game decision-making are deeply intertwined.Based on the analysis of topological structure,the thesis establishes an opinion dynamics model to study the dynamic interaction of human opinions,deeply explores the evolution process of forming the final opinion in a social network,and studies how the individual opinion dynamics shapes the complex group behavior.The thesis is mainly summarized into the following three parts:Firstly,the thesis studies the topology of social networks,and divides the topology into four styles according to different opinions: single leader-follower,multi leader-follower,multi leader-follower-peripheral follower and two-level hierarchical optimization design.Then,the weight of opinions is studied by exploring the influence of each agent to distinguish leaders,followers and peripheral followers.In particular,when the same opinion is adopted by leaders in different network groups,the opinion will be divided into multiple groups.This innovation will open a new perspective for social networks,which can simulate more complex interactive social networks.Secondly,the thesis focuses on the research of multi leader game control.Firstly,we model the social network system and give the optimal synchronization solution of the problem.Then,the situation of agents from local interaction to global coordination is analyzed by simulations more effectively.When the target controller produces different standards for different leaders,different opinions will eventually obtain different solutions.When the target controller produces the same standard for different leaders,the opinion finally obtains a consistent solution.Thirdly,combining the Stackelberg-Nash equilibrium game idea and the cooperative control in the social network system,a two-level hierarchical optimization design scheme,containing the main agent level and the secondary agent level,is established.When solving a problem produces multiple opinions,the system model is established by all opinions of the agent from main agent level and leaders from secondary agent level to obtain the final solution of the preset goal. |
