Global Finite-time Stability for Fractional-order Neural Networks

摘要: This paper is concerned with the global Mittag-Leffler stability (GMLS) and finite-time (GFTS) for fractional Hopfield neural networks (FHNNs) Holder neuron activation functions subject to nonlinear growth. Firstly, four possessing convexity are proposed, which can guarantee that formulas respect derivative established. Correspondingly, a novel principle of convergence in FHNNs developed based on proposed formulas. In addition, by applying Brouwer topological degree theory inequality analysis techniques, proof existence uniqueness equilibrium point addressed. Subsequently, means Lur’e-type Postnikov Lyapunov functional approach, presented finite-time, GMLS GFTS conditions achieved terms linear matrix inequalities (LMIs). Moreover, upper bound setting time calculated accurately. Finally, three numerical examples given verify validity theoretical results.