Yang Yachun, Tu Zhengwen, Wang Liangwei, Cao Jinde, Shi Lei, Qian Wenhua
School of Mathematic and Statistics, Chongqing Three Gorges University, Wanzhou 404100, Chongqing, China.
School of Mathematic and Statistics, Chongqing Three Gorges University, Wanzhou 404100, Chongqing, China.
Neural Netw. 2021 Oct;142:231-237. doi: 10.1016/j.neunet.2021.05.009. Epub 2021 May 11.
This paper investigates H exponential synchronization (ES) of neural networks (NNs) with delay by designing an event-triggered dynamic output feedback controller (ETDOFC). The ETDOFC is flexible in practice since it is applicable to both full order and reduced order dynamic output techniques. Moreover, the event generator reduces the computational burden for the zero-order-hold (ZOH) operator and does not induce sampling delay as many existing event generators do. To obtain less conservative results, the delay-partitioning method is utilized in the Lyapunov-Krasovskii functional (LKF). Synchronization criteria formulated by linear matrix inequalities (LMIs) are established. A simple algorithm is provided to design the control gains of the ETDOFC, which overcomes the difficulty induced by different dimensions of the system parameters. One numerical example is provided to demonstrate the merits of the theoretical analysis.
本文通过设计一种事件触发动态输出反馈控制器（ETDOFC）来研究具有时滞的神经网络（NNs）的H指数同步（ES）。ETDOFC在实际应用中具有灵活性，因为它适用于全阶和降阶动态输出技术。此外，事件发生器减少了零阶保持（ZOH）算子的计算负担，并且不像许多现有事件发生器那样会引起采样延迟。为了获得不太保守的结果，在Lyapunov-Krasovskii泛函（LKF）中采用了时滞分区方法。建立了由线性矩阵不等式（LMIs）制定的同步准则。提供了一种简单的算法来设计ETDOFC的控制增益，该算法克服了系统参数不同维度所带来的困难。给出了一个数值例子来证明理论分析的优点。
