A delay-dependent approach to stability for static recurrent neural networks with mixed time-varying delays

Chien-Yu Lu, Tejen Su, Yihui Su, Shinchun Huang

Research output: Contribution to journalArticle

18 Citations (Scopus)

Abstract

This paper performs a globed stability analysis of a particular class of recurrent neural networks (RNN) in the static neural network models with both discrete and distributed time-varying delays. Both Lipschitz continuous activation function and monotone nondecreasing activation function are considered. Globally delay-dependent stability criteria are derived in the form of linear matrix inequalities (LMI) through the use of Leibniz-Newton formula and relaxation matrices. Moreover, the constraint that derivative of time-varying delays must be smaller than one is released for the proposed control scheme. Finally, two numerical examples are given to illustrate the effectiveness of the proposed criterion.

Original languageEnglish
Pages (from-to)1661-1672
Number of pages12
JournalInternational Journal of Innovative Computing, Information and Control
Volume4
Issue number7
Publication statusPublished - 2008 Jul 1

Fingerprint

Recurrent neural networks
Activation Function
Delay-dependent
Recurrent Neural Networks
Time-varying Delay
Chemical activation
Delay-dependent Criteria
Delay-dependent Stability
Distributed Delay
Monotone Function
Stability criteria
Linear matrix inequalities
Stability Criteria
Neural Network Model
Lipschitz
Matrix Inequality
Linear Inequalities
Stability Analysis
Continuous Function
Derivatives

All Science Journal Classification (ASJC) codes

  • Software
  • Theoretical Computer Science
  • Information Systems
  • Computational Theory and Mathematics

Cite this

@article{dc0a457c2112416f8b01889da265945a,
title = "A delay-dependent approach to stability for static recurrent neural networks with mixed time-varying delays",
abstract = "This paper performs a globed stability analysis of a particular class of recurrent neural networks (RNN) in the static neural network models with both discrete and distributed time-varying delays. Both Lipschitz continuous activation function and monotone nondecreasing activation function are considered. Globally delay-dependent stability criteria are derived in the form of linear matrix inequalities (LMI) through the use of Leibniz-Newton formula and relaxation matrices. Moreover, the constraint that derivative of time-varying delays must be smaller than one is released for the proposed control scheme. Finally, two numerical examples are given to illustrate the effectiveness of the proposed criterion.",
author = "Chien-Yu Lu and Tejen Su and Yihui Su and Shinchun Huang",
year = "2008",
month = "7",
day = "1",
language = "English",
volume = "4",
pages = "1661--1672",
journal = "International Journal of Innovative Computing, Information and Control",
issn = "1349-4198",
publisher = "IJICIC Editorial Office",
number = "7",

}

A delay-dependent approach to stability for static recurrent neural networks with mixed time-varying delays. / Lu, Chien-Yu; Su, Tejen; Su, Yihui; Huang, Shinchun.

In: International Journal of Innovative Computing, Information and Control, Vol. 4, No. 7, 01.07.2008, p. 1661-1672.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A delay-dependent approach to stability for static recurrent neural networks with mixed time-varying delays

AU - Lu, Chien-Yu

AU - Su, Tejen

AU - Su, Yihui

AU - Huang, Shinchun

PY - 2008/7/1

Y1 - 2008/7/1

N2 - This paper performs a globed stability analysis of a particular class of recurrent neural networks (RNN) in the static neural network models with both discrete and distributed time-varying delays. Both Lipschitz continuous activation function and monotone nondecreasing activation function are considered. Globally delay-dependent stability criteria are derived in the form of linear matrix inequalities (LMI) through the use of Leibniz-Newton formula and relaxation matrices. Moreover, the constraint that derivative of time-varying delays must be smaller than one is released for the proposed control scheme. Finally, two numerical examples are given to illustrate the effectiveness of the proposed criterion.

AB - This paper performs a globed stability analysis of a particular class of recurrent neural networks (RNN) in the static neural network models with both discrete and distributed time-varying delays. Both Lipschitz continuous activation function and monotone nondecreasing activation function are considered. Globally delay-dependent stability criteria are derived in the form of linear matrix inequalities (LMI) through the use of Leibniz-Newton formula and relaxation matrices. Moreover, the constraint that derivative of time-varying delays must be smaller than one is released for the proposed control scheme. Finally, two numerical examples are given to illustrate the effectiveness of the proposed criterion.

UR - http://www.scopus.com/inward/record.url?scp=63649086505&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=63649086505&partnerID=8YFLogxK

M3 - Article

AN - SCOPUS:63649086505

VL - 4

SP - 1661

EP - 1672

JO - International Journal of Innovative Computing, Information and Control

JF - International Journal of Innovative Computing, Information and Control

SN - 1349-4198

IS - 7

ER -