Title :
Neural model of electrical devices for circuit simulation
Author :
Dos Santos, Alex L. ; Romariz, Alexandre R S ; De Carvalho, Paulo H P
Author_Institution :
Dept. de Engenharia Eletrica, Brasilia Univ., Brazil
Abstract :
In this work, a neural model for nonlinear circuit devices is proposed. This is an alternative to the modeling of complex ill-defined input-output mapping in new, not well-known structures. The neural model has been used in simulation algorithms revealing good approximation and generalization performance. The model is also been used to estimate output derivatives, which are needed in some simulation and optimization algorithms
Keywords :
circuit analysis computing; circuit optimisation; digital simulation; generalisation (artificial intelligence); multilayer perceptrons; neural nets; nonlinear network analysis; approximation; circuit simulation; generalization performance; neural model; nonlinear circuit devices; optimization algorithms; output derivatives; simulation algorithms; Approximation algorithms; Artificial intelligence; Artificial neural networks; Circuit simulation; Equations; Function approximation; Multilayer perceptrons; Neural networks; Neurons; Nonlinear circuits;
Conference_Titel :
Microwave and Optoelectronics Conference, 1997. Linking to the Next Century. Proceedings., 1997 SBMO/IEEE MTT-S International
Conference_Location :
Natal
Print_ISBN :
0-7803-4165-1
DOI :
10.1109/SBMOMO.1997.646858
