Research On Cluster Synchronization Of Complex Networks And Control For Cluster Formation Of Multi-agent Systems

There are variety of complex networks in the real world,such as transportation network,power transmission network,financial information network and social network,and so on.Due the wide range of applications,complex networks have already received people's concern.The synchronization analysis of complex networks is especially interesting,and has been one of heat issues since the 1990 s.In the scientific researches and social productions,researchers often use a large number of agents with simple structure or simple function to replace the expensive and single system to implement complexity tasks,such as unmanned aerial vehicles and satellite.Multi-agent systems can greatly improve the efficiency.At the same time,the applications of multi-agent systems are based on the theoretical basis of formation control for multi-agent systems.Thus,the researches of formation control for multiagent systems gradually become a frontier subject.In this paper,based on graph theory,Lyapunov stability theory and matrix theory,the problems about cluster synchronization of complex networks and cluster formation of multi-agent systems are studied.The tool of MATLAB is used to verify the correctness of the theory.The main contents of the paper are described as follows.The first chapter introduce the background and the research status of complex networks and multi-agent systems,and gives some basic definitions and preliminaries of the paper.In second chapter,projective cluster synchronization of dynamical networks is considered via pinning impulsive control.Based on the states of driven network,design pinning impulsive control strategy for each cluster of response network.A sufficient condition is given to solve the projective cluster synchronization problem.Finally,the effectiveness of the control mechanism is demonstrated by a numerical example.In chapter three,the cluster formation is considered for nonlinear multi-agent systems based on the event-triggered sampled-data method.Designing event condition and sampling method for each agent by using the sampling information of it's leader agent.A sufficient condition is given to solve the cluster formation problem.Finally,the effectiveness of the control mechanisms are demonstrated by two numerical examples by using the tool of MATLAB.In chapter four,we summary our works and make a prospect for the future work.