Title of article
A New Approach to Train Multilayer Perceptron ANN Using Error Back-propagation and Genetic Algorithms Hybrid: A Case Study of PVTx Estimation of CH4+CF4 Gas Mixture
Author/Authors
Moghadassi، Abdolreza نويسنده Department of Chemical Engineering, Faculty of Engineering, Arak University, Arak, Iran. , , Nikkholgh، Mahmood Reza نويسنده Department of Chemical Engineering, Faculty of Engineering, Arak University, Arak, Iran. , , Hosseini، Sayed Mohsen نويسنده Department of Chemical Engineering, Faculty of Engineering, Arak University, Arak, Iran. , , Parvizian، Fahime نويسنده Department of Chemical Engineering, Faculty of Engineering, Arak University, Arak, Iran. , , Hashemi، Seyyed Jelaladdin نويسنده Petroleum University of Technology, Ahvaz, Iran. ,
Issue Information
فصلنامه با شماره پیاپی 0 سال 2011
Pages
6
From page
177
To page
182
Abstract
A new algorithm to train Multilayer Perceptron Artificial Neural Network using the genetic and Error Back-propagation algorithms Hybrid has been devised. The new algorithm solves the local minimum trap as a natural result of the standard numerical optimization based methods and by following the global minimums the ANN training accuracy has been highly improved. There are many algorithms for training a Multilayer Perceptron ANN to estimate the PVTx of CH4+CF4 gas mixture. The new devised algorithm is compared and evaluated against these algorithms and indicates a better accuracy.
Journal title
International Journal of Industrial Chemistry (IJIC)
Serial Year
2011
Journal title
International Journal of Industrial Chemistry (IJIC)
Record number
655153
Link To Document