Use of genetic algorithms with back propagation in training of feed-forward neural networks

Michael McInerney; Atam P. Dhawan

Use of genetic algorithms with back propagation in training of feed-forward neural networks

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The problem of genetic algorithms is that they are inherently slow. In this paper we present a hybrid of genetic and back-propagation algorithms (GA-BP) which should always find the correct global minima without getting stuck at local minima. Various versions of the GA-BP method are presented and experimental results show that GA-BP algorithms are as fast as the back-propagation algorithm and do not get stuck at local minima. The proposed GA-BP algorithms are also not sensitive to the values of momentum and learning rate used in back-propagation and can be made independent of the learning rate and momentum. It is shown that the adaptive GA-BP algorithm can provide the optimal learning rate and better performance than simple back-propagation.

Original language	English (US)
Title of host publication	1993 IEEE International Conference on Neural Networks
Editors	Anon
Publisher	Publ by IEEE
Pages	203-208
Number of pages	6
ISBN (Print)	0780312007
State	Published - 1993
Externally published	Yes
Event	1993 IEEE International Conference on Neural Networks - San Francisco, CA, USA Duration: Mar 28 1993 → Apr 1 1993

Publication series

Name	1993 IEEE International Conference on Neural Networks

Other

Other	1993 IEEE International Conference on Neural Networks
City	San Francisco, CA, USA
Period	3/28/93 → 4/1/93

All Science Journal Classification (ASJC) codes

General Engineering
Control and Systems Engineering
Software
Artificial Intelligence

Cite this

@inproceedings{1756268807b94034a92b47e6a6beb238,

title = "Use of genetic algorithms with back propagation in training of feed-forward neural networks",

abstract = "The problem of genetic algorithms is that they are inherently slow. In this paper we present a hybrid of genetic and back-propagation algorithms (GA-BP) which should always find the correct global minima without getting stuck at local minima. Various versions of the GA-BP method are presented and experimental results show that GA-BP algorithms are as fast as the back-propagation algorithm and do not get stuck at local minima. The proposed GA-BP algorithms are also not sensitive to the values of momentum and learning rate used in back-propagation and can be made independent of the learning rate and momentum. It is shown that the adaptive GA-BP algorithm can provide the optimal learning rate and better performance than simple back-propagation.",

author = "Michael McInerney and Dhawan, {Atam P.}",

year = "1993",

language = "English (US)",

isbn = "0780312007",

series = "1993 IEEE International Conference on Neural Networks",

publisher = "Publ by IEEE",

pages = "203--208",

editor = "Anon",

booktitle = "1993 IEEE International Conference on Neural Networks",

note = "1993 IEEE International Conference on Neural Networks ; Conference date: 28-03-1993 Through 01-04-1993",

}

Use of genetic algorithms with back propagation in training of feed-forward neural networks. / McInerney, Michael; Dhawan, Atam P.
1993 IEEE International Conference on Neural Networks. ed. / Anon. Publ by IEEE, 1993. p. 203-208 (1993 IEEE International Conference on Neural Networks).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

TY - GEN

T1 - Use of genetic algorithms with back propagation in training of feed-forward neural networks

AU - McInerney, Michael

AU - Dhawan, Atam P.

PY - 1993

Y1 - 1993

N2 - The problem of genetic algorithms is that they are inherently slow. In this paper we present a hybrid of genetic and back-propagation algorithms (GA-BP) which should always find the correct global minima without getting stuck at local minima. Various versions of the GA-BP method are presented and experimental results show that GA-BP algorithms are as fast as the back-propagation algorithm and do not get stuck at local minima. The proposed GA-BP algorithms are also not sensitive to the values of momentum and learning rate used in back-propagation and can be made independent of the learning rate and momentum. It is shown that the adaptive GA-BP algorithm can provide the optimal learning rate and better performance than simple back-propagation.

AB - The problem of genetic algorithms is that they are inherently slow. In this paper we present a hybrid of genetic and back-propagation algorithms (GA-BP) which should always find the correct global minima without getting stuck at local minima. Various versions of the GA-BP method are presented and experimental results show that GA-BP algorithms are as fast as the back-propagation algorithm and do not get stuck at local minima. The proposed GA-BP algorithms are also not sensitive to the values of momentum and learning rate used in back-propagation and can be made independent of the learning rate and momentum. It is shown that the adaptive GA-BP algorithm can provide the optimal learning rate and better performance than simple back-propagation.

UR - http://www.scopus.com/inward/record.url?scp=84943266163&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84943266163&partnerID=8YFLogxK

M3 - Conference contribution

AN - SCOPUS:84943266163

SN - 0780312007

T3 - 1993 IEEE International Conference on Neural Networks

SP - 203

EP - 208

BT - 1993 IEEE International Conference on Neural Networks

A2 - Anon, null

PB - Publ by IEEE

T2 - 1993 IEEE International Conference on Neural Networks

Y2 - 28 March 1993 through 1 April 1993

ER -

Use of genetic algorithms with back propagation in training of feed-forward neural networks

Abstract

Publication series

Other

All Science Journal Classification (ASJC) codes

Other files and links

Fingerprint

Cite this