Title :
A neural-network approach to modeling and analysis
Author :
Chen, Chen-Yuan ; Chen, Cheng-Wu ; Chiang, Wei-Ling ; Hwang, Jing-Dong
Author_Institution :
Dept. of Marine Environ. & Eng., Nat. Sun Yat-Sen Univ., Kaohsiung, Taiwan
Abstract :
A backpropagation network can always be used in modeling. This study is concerned with the stability problem of a neural network (NN) system which consists of a few subsystems represented by NN models. In this paper, the dynamics of each NN model is converted into linear inclusion representation. Subsequently, based on the representations, the stability conditions in terms of Lyapunov´s direct method is derived to guarantee the asymptotic stability of NN systems.
Keywords :
Lyapunov methods; asymptotic stability; backpropagation; digital simulation; neural nets; stability criteria; Lyapunov direct method; NN model dynamics; asymptotic stability; backpropagation network; linear inclusion representation; modeling; neural network system stability; neural-network approach; stability; subsystems; Aerodynamics; Asymptotic stability; Biological neural networks; Biological system modeling; Interconnected systems; Large-scale systems; Neural networks; Nonlinear control systems; Stability analysis; Stability criteria;
Conference_Titel :
Tools with Artificial Intelligence, 2002. (ICTAI 2002). Proceedings. 14th IEEE International Conference on
Print_ISBN :
0-7695-1849-4
DOI :
10.1109/TAI.2002.1180843
