Global Exponential Periodicity of a Class of Neural Networks with Recent-History Distributed Delays

XF Yang,XF Liao,GM Megson,DJ Evans
DOI: https://doi.org/10.1016/j.chaos.2004.11.014
IF: 9.922
2005-01-01
Chaos Solitons & Fractals
Abstract:In this paper, we propose to study a class of neural networks with recent-history distributed delays. A sufficient condition is derived for the global exponential periodicity of the proposed neural networks, which has the advantage that it assumes neither the differentiability nor monotonicity of the activation function of each neuron nor the symmetry of the feedback matrix or delayed feedback matrix. Our criterion is shown to be valid by applying it to an illustrative system.
What problem does this paper attempt to address?