Title :
Stabilization of stochastic recurrent neural networks
Author :
Sanchez, Edgar N. ; Perez, Jose P.
Author_Institution :
CINVESTAV, Unidad Guadalajara, Mexico
Abstract :
The paper presents the stabilization of a dynamic neural network disturbed by additive Gaussian noise. This stabilization is achieved using a quadratic Lyapunov function. A control law is derived, which ensures that the neural network state becomes globally uniform ultimately bounded (GUUB) in probability. The applicability of the proposed control law is illustrated by an example.
Keywords :
Gaussian noise; Lyapunov methods; control system synthesis; neurocontrollers; nonlinear control systems; recurrent neural nets; stability; stochastic systems; additive Gaussian noise; control design; control law; dynamic neural network; globally uniform ultimately bounded neural network state; nonlinear systems; quadratic Lyapunov function; stochastic control Lyapunov function; stochastic recurrent neural network stabilization; Additive noise; Associative memory; Electronic mail; Lyapunov method; Neural networks; Nonlinear systems; Recurrent neural networks; Stability analysis; Stochastic processes; Stochastic resonance;
Conference_Titel :
Intelligent Control, 2002. Proceedings of the 2002 IEEE International Symposium on
Print_ISBN :
0-7803-7620-X
DOI :
10.1109/ISIC.2002.1157804
