Finite-time and fixed-time anti-synchronization of neural networks with time-varying delays

In this paper, the finite-time and fixed-time anti-synchronization of master-slave dynamical systems with time-varying delays are investigated. The feedback controller is designed only depending on the system state at present time t, but independent of the delayed states, which would be much easier to be verified and realized in practice. Rigorous analysis is developed in two cases with respect to the different ranges of initial state of error systems. Then the absolute value of each error state component can be considered as that, flowing from the initial value to 1 first, then from 1 to 0, and the time it needs in this whole process is finite. As special cases, several neural network models with unbounded time delays are addressed to illustrate the effectiveness and efficiency of our obtained results. (C) 2018 Elsevier B.V. All rights reserved.