Direct solution method for finite element analysis using Hopfield neural network

Author

Yamashita, Hideo ; Kowata, Norio ; Cingoski, V. ; Kaneda, Kazufumi

Author_Institution

Fac. of Eng., Hiroshima Univ., Japan

Volume

31

Issue

3

fYear

1995

fDate

5/1/1995 12:00:00 AM

Firstpage

1964

Lastpage

1967

Abstract

One property of the Hopfield neural network is the monotonous minimization of energy as time proceeds. In this paper, this property is applied to minimize the energy functional obtained by ordinary finite element analysis. The mathematical representation and correlation between finite element and neural network calculus are presented. The selection of the sigmoid function and its influence on the iteration process is discussed. The obtained results using the proposed method show excellent agreement with theoretical solutions

Keywords

Hopfield neural nets; electromagnetic fields; finite element analysis; iterative methods; EM field problems; Hopfield neural network; energy functional; finite element analysis; iteration process; mathematical representation; monotonous minimization; neural network calculus; sigmoid function; Artificial neural networks; Biological neural networks; Calculus; Electromagnetic fields; Electrostatics; Finite element methods; Hopfield neural networks; Neural networks; Neurons; Power engineering and energy;

fLanguage

English

Journal_Title

Magnetics, IEEE Transactions on

Publisher

ieee

ISSN

0018-9464

Type

jour

DOI

10.1109/20.376426

Filename

376426