Synchronization of competitive neural networks with different time scales

In this paper, the exponential synchronization problem for a class of competitive neural networks is investigated. Moreover, without assuming the active functions to be differentiable and bounded, some exponential synchronization criteria are devised by Lyapunov functionals, linear matrix inequality (LMI) approach, the Leibniz–Newton formula. The conditions are less conservative than existing ones reported in the literature for delayed neural networks. The applicability of the conditions are validated by an illustrative example.