Fixed-time synchronization of delayed inertial memristive neural networks under denial-of-service attacks.

Abstract

This paper investigates the fixed-time drive-response synchronization problem of delayed inertial memristive neural networks under denial-of-service (DoS) attacks. To address the complex dynamics arising from inertial effects, time-varying delays, uncertainties in memristive switching, and the interplay of intermittent DoS attacks, a fixed-time control scheme is proposed based on fixed-time stability theory. By constructing an appropriate Lyapunov function, several sufficient conditions are derived to guarantee synchronization within a fixed-time under DoS attack scenarios. Furthermore, an explicit expression for the upper bound of the settling time is provided, quantitatively revealing the dependence of settling time on controller parameters and the intensity of DoS attacks. The impact of attack intensity on synchronization performance is also analyzed. Notably, the proposed theoretical framework applies not only to aperiodic DoS attacks but also covers periodic DoS attacks, and remains effective even under normal operating conditions without DoS attacks, thereby significantly broadening the applicability and generality of the results. Finally, numerical simulations are presented to further validate the effectiveness and robustness of the proposed control strategy.