New LMI constraint-based settling-time estimation for finite-time stability of fractional-order neural networks

Abstract This study aims to analyze the finite-time stability performance of both non-delayed and delayed fractional-order neural networks. Our primary aim is to investigate the finite-time stability characteristics by introducing a novel inequality designed for estimating the settling time. This fresh inequality serves as the foundation for establishing sufficient criteria, formulated as linear matrix inequalities, which guarantee the finite-time stability of both non-delayed and delayed fractional-order neural networks. Additionally, we underscore the importance of incorporating comprehensive information regarding the lower and upper bounds of the activation function, especially in the context of the proposed non-delayed model. Unlike the previous works, in this paper, the linear matrix inequality technique has been adopted towards the finite-time stability behavior of the proposed model. At last, some numerical examples are examined to validate the efficacy and conservatism of the presented approach and established theoretical results over the existing literature.