一类四元数值Cohen-Grossberg神经网络的固定时间同步  

Fixed-time Synchronization for a Class of Quaternion-valued Cohen-Grossberg Neural Networks

作　　者：陈星宇 朱培勇[1] CHEN Xingyu;ZHU Peiyong(School of Mathematical Science,University of Electronic Science and Technology of China,Chengdu 611731,China)

机构地区：[1]电子科技大学数学科学学院,成都611731

出　　处：《四川轻化工大学学报（自然科学版）》2023年第1期93-100,共8页Journal of Sichuan University of Science & Engineering(Natural Science Edition)

基　　金：国家自然科学基金项目(61703354)。

摘　　要：研究一类具有时变时滞的四元数Cohen-Grossberg神经网络(QVCGNNs)的固定时间同步问题。首先,在使用直接法把四元数神经网络当作一个整体的前提下,利用四元数符号函数的性质设计了新颖有效的四元数值控制器使得驱动响应系统实现固定时间同步;然后,通过构造适当的Lyapunov-Krasovskii泛函,运用不等式技巧得到了保证四元数Cohen-Grossberg神经网络驱动响应系统固定时间同步的判据;最后,提供了一个数值仿真算例用以验证所提出方法的可靠性。The fixed-time synchronization problem for a class of quaternion-valued Cohen-Grossberg neural networks(QVCGNNs)with time-varying delays is studied.Firstly,on the premise of using the direct method to treat the quaternion-valued neural networks as an entirety without any decomposition,based on the nature of the sign function related to quaternion,one simple and effective quaternion controller is designed to guarantee drive system and response system achieved synchronization within a fixed time.Next,by constructing the appropriate Lyapunov-Krasovskii function and utilizing inequalities skills,the criterion for ensuring the fixed time synchronization of the quaternion Cohen-Grossberg neural network drive response system is obtained.Finally,one numerical example is given to show the reliability and validity of our theoretical analyses.

关 键 词：四元数神经网络 固定时间同步 时变时滞 

分 类 号：O175[理学—数学]