| Modern network technology is closely related to social production and development,which can make human life more convenient.However,with the development of modern computer technology,the complexity of network and the accuracy required for research increase rapidly,which requires researchers to analyze and study complex network more deeply and comprehensively,so that the theoretical knowledge can solve the practical problems in production activities and social life more accurately.Multi-agent system is an important research in the complex network.It is derived from the theoretical basis of artificial intelligence,distributed control,distributed computing and so on.It can well describe the interaction between agents in the system,and is closely concerned by scholars from all circles.It can be used to study UAV formation,robot control and so on.In the research of multi-agent system,the problem of flocking is the main research object.Flocking mainly describes the orderliness of agents in the system,which is mainly reflected in the synchronization of movement modes among each agent,such as having the same speed,movement direction,temperature and so on.In this paper,based on the optimization and transformation of the thermomechanical Cucker-Smale model,the speed of each agent has a certain free will,and the continuous and discrete thermomechanical Cucker-Smale model is constructed to increase the speed disturbance,so that the optimized model can be more accurate and practical.By using graph theory,matrix theory,differential equation and other related knowledge,the sufficient conditions of the flocking are obtained.In the case of flocking,the speed and temperature of each agent tend to a fixed value,and their spatial position tends to a stable shape.After the rigorous theoretical deduction,it is concluded that when the initial value,parameters and speed disturbance of the system satisfies certain conditions,the continuous and discrete models can achieve flocking,but the initial value and parameters of the continuous and discrete models need to satisfie some different conditions. |