Research On Opinion Evolution Model Based On The Flocking Policy On The Internet

With the development of internet technology, the impact of network information on the real-life is gradually increasing. In order to reduce the negative impact of information and better use of the internet to people's lives, the evolution of public opinion research on the Internet has become a hot topic. It is a cross-disciplinary subject involving a number of areas, including social psychology, social physics, information technology and computer science. The classical opinon evolution model focused on the nonlinear characteristics of the process, and the Bounded Confidence model is the most widely used one. In this work I propose a new model to adapt to the complex network environment, and use the simulation experimental data to analyse the model.The Internet technology breaks the limit of time and space, and the social networking services reflect the real interpersonal networks on the Internet. The truthful information of personal accelerates the evolution of the public opinion. I proposed a multi-agent system model to explain the process of opinion evolution in cyberspace. The location realationship between agents in the two-dimensional space means the interpersonal in the real-life. The "eyeshot" is used to explain the individual's ability to obtain information in cyberspace. The communaction between agtents is limited by the eyeshot and agents adjust their opinion according to the Bounded Confidence Algorithm. With the simulation result, we find that the new model effectively ingerited the basic characteristics of the H&K's Bounded Confidence model. On this basis, I investigated the opinion dynamics of the new system with different composition of the population. The agents with the larger eyeshot will accelerate the formation of public opinion; the agents with tiny bounded confidence and extreme opinion will induce final opinion to deflection; the agents with all the above characteristics will make the system to the polarization.The dynamic network of relationships is a classic social network structure. In this study of opinion dynamics, I use the Boid model to simulate the social network structure and focus basically on the influence of the agent's behavior in Cyberspace. In Cyberspace, there are two psychological elements of individual behaviors in the interaction between people, curiosity and wariness. People are curious about each other when they interact, and wary of neighbors in order to protect their privacy at the same time. In this work, eyeshot and the attractive-repulsive radius are used to limit the communication area between agents and to define the range in which agents can interact as well. Through simulation, it is found that the attraction impacts not only on the agent's motion but also upon the opinion dynamics. The eyeshot is a restriction condition which can eliminate the effects. In addition, heterogeneous initial conditions are used instead of homogeneous distribution for a more realistic simulation. I introduce different types of agents to the population, i. e. the leader agent. It is showed in the results that the leader agent can delay the convergence process and change the probability of different opinion dynamics occurred with extreme opinions.The topic transimission is an important part of the process in the opinion evolution. In this work, I use the virus model to simulate the process of topic transimission and define the value of "trigger" to explain the individual degree of curiosity for the unknown event. The analysis is focused on the opinion dynamics of system with different trigger. The result shows that there is no difference when the agent's trigger is small but some new characteristics with the large agent's trigger. When the topic transimission is not synchronized with the evolution of public opinion, the edge opinions will survive in the system. The size of edge opinions will strengthen with increasing eyeshot and weaken with the increasing attractive radius. On this basis, with the improving algorithm, I investigate the impact of opinion leaders on the process of topic transimission and opinion evolution. The system evolves from inertia state to the active state with the increasing eyeshot and attractive radius. And the final opinions will converge to the two opposing views.By the study and analysis, finally, I proposed a more complete model of topic transimission and opinion evolution and evaluate the impact of different interventions on the evolution of public opinion.