Premium
Existence and global exponential stability of almost periodic solution for delayed competitive neural networks with discontinuous activations
Author(s) -
Tan Yanxiang,
Jing Ke
Publication year - 2016
Publication title -
mathematical methods in the applied sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.719
H-Index - 65
eISSN - 1099-1476
pISSN - 0170-4214
DOI - 10.1002/mma.3732
Subject(s) - mathematics , exponential stability , artificial neural network , stability (learning theory) , class (philosophy) , lyapunov function , work (physics) , control theory (sociology) , exponential function , exponential growth , mathematical analysis , computer science , nonlinear system , artificial intelligence , control (management) , mechanical engineering , physics , quantum mechanics , machine learning , engineering
In this paper, we study a class of delayed competitive neural networks with discontinuous activations. Without assuming the boundedness and local Lipschizian on the activation functions, some new criteria ensuring the existence and global exponential stability of almost periodic solutions for the neural network model considered in this work are established by constructing some suitable Lyapunov functionals and employing the theory of nonsmooth analysis. Finally, we present some applications and numerical examples with simulations to show the effectiveness of our main results. Copyright © 2016 John Wiley & Sons, Ltd.