Predefined‐time synchronization of coupled inertial complex‐valued neural networks based on exponential scaling function

Abstract This paper investigates the predefined‐time synchronization of coupled inertial complex‐valued neural networks. By employing variable transformation and real‐imaginary separation methods, the coupled inertial complex‐valued neural networks are transformed equivalently to higher dimensional systems containing only first‐order derivatives. An effective controller with a time‐dependent variant of the exponential function is designed and sufficient conditions for achieving predefined‐time synchronization of coupled inertial complex‐valued neural networks are derived. This paper successfully extends existing research on predefined‐time synchronization from driven‐response inertial complex‐valued neural networks to a general class of coupled inertial complex‐valued neural networks. Moreover, unlike the sign functions or power functions commonly used in existing studies of predefined‐time synchronization, this paper introduces a new time‐dependent variant of the exponential function in the controller design. Finally, the validity of the proposed results is verified through a numerical example.