Weighted Pseudo Almost-Automorphic Solutions of Quaternion-Valued RNNs With Mixed Delays

This work deals with a nonlinear differential equation for a quaternion-valued recurrent neural network. By using the contraction mapping principle and some differential inequalities, we directly studied the existence and the global exponential stability of weighted pseudo-almost automorphic solution for this class of quaternion-valued neural networks. Here, methods were applied without a real or a complex decomposition of the equation system. In addition, an application verifying our results and its numerical simulation was given. The generated results about the weighted pseudo-almost automorphic solution of the considered model are new.

Automated summary

This work focuses on a nonlinear differential equation for a quaternion-valued recurrent neural network. The existence and global exponential stability of a weighted pseudo-almost automorphic solution for this type of neural network is directly studied using the contraction mapping principle and some differential inequalities. The methods used in this study do not involve a real or complex decomposition of the equation system. Additionally, an application and numerical simulation are provided to verify the results obtained. The generated results about the weighted pseudo-almost automorphic solution in the considered model are novel.