Title :
Stability analysis of recurrent neural networks - a Volterra integro-differential equation approach
Author :
Liu, Pingzhou ; Han, Qing-Long
Author_Institution :
Fac. of Informatics & Commun., Central Queensland Univ., Rockhampton, Qld., Australia
Abstract :
The stability of a special class of nonlinear Volterra integro-differential systems are analyzed by comparing them to linear Volterra integro-differential systems. The results are used to determine the stability properties of recurrent neural networks with distributed delays, including constant discrete delays as a special case. The obtained stability criteria have unified and extended many existing results on recurrent neural networks.
Keywords :
Volterra equations; asymptotic stability; delays; integro-differential equations; nonlinear equations; recurrent neural nets; Volterra integro-differential equation approach; constant discrete delays; distributed delays; nonlinear Volterra integro-differential systems; recurrent neural networks; stability analysis; stability criteria; Biological neural networks; Delay effects; Integrodifferential equations; Neural networks; Neurodynamics; Neurons; Recurrent neural networks; Stability analysis; Stability criteria; Symmetric matrices;
Conference_Titel :
Decision and Control, 2004. CDC. 43rd IEEE Conference on
Print_ISBN :
0-7803-8682-5
DOI :
10.1109/CDC.2004.1429669
