This manuscript mainly deals with quaternion‐valued neural networks (QVNNs) with delays and inertial term. Using Wirtinger inequality and coincidence degree theory, a new sufficient criterion to ensure the existence of antiperiodic solution of involved quaternion‐valued neural networks is derived. With the aid of Lyapunov function, we discuss the exponential stability of antiperiodic solutions to quaternion‐valued neural networks. Numerical simulations are presented to illustrate the established theoretical findings.

中文翻译：

---

延迟惯性四元数值神经网络的反周期解

该手稿主要处理具有时滞和惯性项的四元数神经网络（QVNN）。利用维特林格不等式和重合度理论，推导了一个新的充分准则，以确保所涉及的四元数值神经网络的反周期解的存在。借助Lyapunov函数，我们讨论了四元数值神经网络的反周期解的指数稳定性。数值模拟表明了已建立的理论发现。