Improved Stability Analysis of Neural Networks with Time Delay Based on Variable Augmented Free Weight Matrix

Abstract This article investigates the stability of time-delay neural network systems by the free-weighting matrices based on variable- augmentation.In the process of accurately evaluating the stability of delayed neural network systems, the conservative simplification of stability criteria has received widespread attention. In this paper, a variable-augmentation-based free-weighting matrix method is applied to time-varying delayed neural network systems, and unnecessary free weighting matrices are removed. Some results with fewer free-weighting matrices are obtained, while maintaining their stability and conservatism. Finally, a specific neural network numerical example was used to demonstrate the effectiveness of this method.