A Novel Approach to Simulate Lane-Emden and Emden-Fowler Equations Using Curriculum Learning-Based Unsupervised Symplectic Artificial Neural Network

A Novel Approach to Simulate Lane-Emden and Emden-Fowler Equations Using Curriculum Learning-Based Unsupervised Symplectic Artificial Neural Network

Year:    2023

Author:    Arup Kumar Sahoo, S. Chakraverty

East Asian Journal on Applied Mathematics, Vol. 13 (2023), Iss. 2 : pp. 276–298

Abstract

This paper investigates the impact of the curriculum learning process in a multilayer neural network (NN) for solving the Lane-Emden and Emden-Fowler models. Starting from the training of a neural network in a small domain, we gradually expanded the domain. The symplectic NN trial solution is used for solving titled models. Feedforward NN and error back-propagation algorithms are used to minimize the error function and modify the parameters. The consistency of the algorithm is demonstrated by solving several problems. Calculating different types of errors (MSE up to 1E-10), we show an excellent agreement between the current simulations and existing results.

You do not have full access to this article.

Already a Subscriber? Sign in as an individual or via your institution

Journal Article Details

Publisher Name:    Global Science Press

Language:    English

DOI:    https://doi.org/10.4208/eajam.2022-115.300922

East Asian Journal on Applied Mathematics, Vol. 13 (2023), Iss. 2 : pp. 276–298

Published online:    2023-01

AMS Subject Headings:   

Copyright:    COPYRIGHT: © Global Science Press

Pages:    23

Keywords:    Curriculum learning symplectic neural network unsupervised Lane-Emden equation Emden-Fowler equation.

Author Details

Arup Kumar Sahoo

S. Chakraverty