Stability analysis of delayed neural networks via novel delay-dependent LKF and integral inequality

The current paper is concerned with the stability analysis of delayed neural networks. In the case that the delay derivative is restricted with an upper bound only, the augmented LKFs often contain high-degree terms of the time-varying delay, resulting in the non-convex derivatives of LKFs, which can be solved by introducing extra delay-multiplied state variables to transform the non-convex delay-dependent terms into convex ones. To make fuller use of the delay-multiplied state variables and the delay-derivative-dependent information, these delay-multiplied state variables are introduced into an LKF and the integral inequality through the proper augmentation in this paper. Meanwhile, some free-matrix-based zero equations are introduced into this delay-dependent inequality to provide more freedom. By applying the augmented LKF and the novel integral inequality, a delay-dependent stability criterion of delayed neural networks with less conservatism is established, whose advantages are verified by three examples.