Stabilization of complex-valued neural networks subject to semi-Markov jumping parameters: A dynamic event-triggered protocol

For continuous-time complex-valued neural networks, this paper addresses the state-feedback stabilization issue via dynamic event-triggered protocol. Aiming at random parameters' switching, semi-Markov jump model surpasses the Markov jump model in terms of its generality, enabling us to effectively capture the occurrence of random abrupt alterations in both the structure and parameters of complex-valued neural networks. To optimize packet transmission, a new dynamic event-based protocol is introduced to judge whether the previous signal transmission continues. The design of this protocol takes into full consideration the imaginary part characteristics of the system, while also integrating the system modes and dynamic variables. Utilizing an appropriate Lyapunov functional that contains auxiliary internal dynamical variables, the desired stability is proposed. Eventually, the effectiveness of theoretical findings is ultimately validated through two numerical simulations.