On The Synchronization And Consensus Of Complex Network Systems

The synchronization and consensus of complex network systems have received considerable attention due to their widely potential applications in various areas such as information science,social science,engineering science and so on.In this thesis,we focus on the problems of synchronization for coupled neural networks(CNNs)and consensus for multi-agent systems(MAS).In detail,we investigate the synchronization for CNNs with time-varying delay,the average consensus for first-order MAS under directed communication networks with balanced and strongly connected topology and the finite-time leader-following lag consensus for second-order MAS.Our work is divided into the following aspects:(1)The synchronization problem for a class of neural networks with time-varying delay and hybrid couplings is discussed.By using a Lyapunov-Krasovskii functional(LKF)containing a triple integral term,applying the theory of Kronecker product of matrices,the Wirtinger-based inequality,the method of reciprocally convex combination and the technique of linear matrix inequality(LMI)to estimate the derivative of the LKF,the delay-dependent synchronization criteria in terms of LMIs are newly obtained.(2)The problem of average consensus for MAS with time-varying delay and switching communication topology is investigated.By constructing a new LKF and using the Wirtinger-based inequality and reciprocally convex combination approach to handle the integral terms of the derivative of the LKF,a novel delay-dependent criterion for average consensus in terms of LMIs is obtained.(3)The problem of finite-time leader-following lag consensus for second-order MAS is solved.The disturbances of follower agents are assumed to generated by the exogenous systems.By using the disturbance-observer-based control and LMI approach,the disturbance observer is developed to ensure the stability of the disturbance error system.Then by utilizing the nonsingular fast terminal sliding mode(NFTSM)control technique,the corresponding NFTSM controller is designed to guarantee the finite-time lag consensusability between the follower agents and leader agent.