| As a new type of complex network, switched neural networks have been introduced into the theory of control. Switched neural networks has received great attention of researchers as its widely use in practical applications and theoretical value. Due to the interaction among the impulsive phenomenon and time delay, the dynamic behavior of switched neural networks becomes more complicated and the stability of such systems deserves a further study. In addition, impulses can be used as a controller to stabilize systems because it can change the system state instantaneously. The synchronization problem of the master-slave system by using impulsive controller has also become a major research direction. The main results of the thesis can be summarized as follows:1. The stability of switched neural networks subject to impulsive effects and time-varying delay is investigated in this section. The average dwell time method are employed in the case that all subsystems are stable. The state-dependent switching rule is designed to divide the whole state space into several subspaces, if there exists unstable subsystem. Through the above two methods, sufficient conditions in terms of linear matrix inequalities(LMIs) are derived via the multiple Lyapunov functions to guarantee the stability of the described system, and they can be solved with MATLAB tools.2. The synchronicity analysis of drive-response switched neural networks with impulsive controller is studied. In this part, two kinds of impulsive controllers are designed. One is based on full-dimension subsystem’s state. The other is a part-dimension impulsive controller which is based on the interconnection among the subsystems of the neural networks. Serval sufficient conditions for the synchronization of drive-response switched neural networks are derived by the dwell time approach. The conservatism of the results can be reduced through designing a novel segmented time-varying Lyapunov function. |
PDF Full Text Request