Fixed-time synchronization of competitive neural networks with multiple time-scale - Archive ouverte HAL Accéder directement au contenu
Article Dans Une Revue IEEE Transactions on Neural Networks and Learning Systems Année : 2021

Fixed-time synchronization of competitive neural networks with multiple time-scale

Résumé

In this brief we investigate the fixed-time synchronization of competitive neural networks with multiple timescales. These neural networks play an important role in visual processing, pattern recognition, neural computing, etc. Our main contribution is the design of a novel synchronizing controller which does not depend on the ratio between the fast and slow time scales. This feature makes the controller easy to implement since it is designed through well-posed algebraic conditions (i.e. even when the ratio between the time scales goes to 0 the controller gain is well defined and does not go to infinity). Last but not least, the closed-loop dynamics is characterized by a high convergence speed with a settling time which is upper-bounded and the bound is independent of the initial conditions. A numerical simulation illustrates our results and emphasizes their effectiveness. Index Terms-Competitive neural network, multiple timescale feature, fixed-time synchronization, continuous control method.

Domaines

Automatique
Fichier principal
Vignette du fichier
TNNLS-2020.pdf (593.41 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-03455236 , version 1 (29-11-2021)

Identifiants

Citer

Wu Yang, Yan-Wu Wang, Irinel-Constantin Morarescu, Xiao-Kang Liu, Yuehua Huang. Fixed-time synchronization of competitive neural networks with multiple time-scale. IEEE Transactions on Neural Networks and Learning Systems, 2021, pp.Early Access. ⟨10.1109/TNNLS.2021.3052868⟩. ⟨hal-03455236⟩
17 Consultations
103 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More