Stability of Quasi-Periodic Orbits in Recurrent Neural Networks
✍ Scribed by R. L. Marichal; J. D. Piñeiro; E. J. González; J. Torres
- Publisher
- Springer US
- Year
- 2010
- Tongue
- English
- Weight
- 401 KB
- Volume
- 31
- Category
- Article
- ISSN
- 1370-4621
No coin nor oath required. For personal study only.
📜 SIMILAR VOLUMES
In this paper we introduce recurrent dimensions of discrete dynamical systems and we give upper and lower bounds of the recurrent dimension of the quasi-periodic orbits. We show that these bounds have different values according to the algebraic properties of the frequency and we investigate these di
## Abstract This paper proposes a class of more general model of recurrent neural networks with __functional__ delay, which has been found more suitable to apply directly. Simple and easily checkable conditions of existence, uniqueness, and global exponential stability of periodic solution for the
we consider a nonlinear discrete-time system 4n + 1) = k(n) + g(y(n -k)), Y(n + 1) = Dy(n) -g(4n -k)), n E N, arising as a discrete-time network of two neurons with McCulloch-Pitts nonlinearity, where p E (0, l), k is a positive integer, and g is a signal transmission function with a threshold 6. W
In this paper recurrent neural networks with time-varying delays and continuously distributed delays are considered. Sufficient conditions for the existence and exponential stability of the anti-periodic solutions are established, which are new and complement previously known results.