LAUSR

As is known to all that the Lyapunov‐Krasovskii functional (LKF) method plays a significant role in deriving exponential stability criteria of neural networks with a time‐varying delay. However, when the LKF method is adopted, the condition that a functional is required for a neural network with a delay varying in a delay interval is so strong that it may be hard to be satisfied and lead to a conservative criterion. Therefore, a generalized LKF method is proposed by weakening the strong condition in this paper. Then, new exponential stability criteria are derived via applying the proposed method. Finally, the effectiveness of the derived criteria is verified by two numerical examples.