Using Differential Evolution for GEP Constant
Creation
Q. Zhang, C. Zhou, W. Xiao, P. Nelson, and X. Li. Late breaking paper at Genetic and Evolutionary
Computation Conference (GECCO'2006), Seattle, WA, USA, (8-12 July 2006)
Abstract
Gene Expression Programming (GEP) is a new
evolutionary algorithm that incorporates both the idea
of simple, linear chromosomes of fixed length used in
Genetic Algorithms (GAs) and the structure of different
sizes and shapes used in Genetic Programming (GP). As
with other genetic programming algorithms, GEP has
difficulty finding appropriate numeric constants for
terminal nodes in the expression trees. In this paper,
we describe a new approach of constant generation using
Differential Evolution (DE), which is a simple
real-valued GA that has proven to be robust and
efficient on parameter optimisation problems. Our
experimental results on two symbolic regression
problems show that the approach significantly improves
the performance of the GEP algorithm. The proposed
approach can be easily extended to other Genetic
Programming variants.
%0 Conference Paper
%1 Zhang:gecco06lbp
%A Zhang, Qiongyun
%A Zhou, Chi
%A Xiao, Weimin
%A Nelson, Peter C.
%A Li, Xin
%B Late breaking paper at Genetic and Evolutionary
Computation Conference (GECCO'2006)
%C Seattle, WA, USA
%D 2006
%E Grahl, Jörn
%K DE algorithms, expression gene genetic programming,
%T Using Differential Evolution for GEP Constant
Creation
%U http://www.cs.bham.ac.uk/~wbl/biblio/gecco2006etc/papers/lbp130.pdf
%X Gene Expression Programming (GEP) is a new
evolutionary algorithm that incorporates both the idea
of simple, linear chromosomes of fixed length used in
Genetic Algorithms (GAs) and the structure of different
sizes and shapes used in Genetic Programming (GP). As
with other genetic programming algorithms, GEP has
difficulty finding appropriate numeric constants for
terminal nodes in the expression trees. In this paper,
we describe a new approach of constant generation using
Differential Evolution (DE), which is a simple
real-valued GA that has proven to be robust and
efficient on parameter optimisation problems. Our
experimental results on two symbolic regression
problems show that the approach significantly improves
the performance of the GEP algorithm. The proposed
approach can be easily extended to other Genetic
Programming variants.
@inproceedings{Zhang:gecco06lbp,
abstract = {Gene Expression Programming (GEP) is a new
evolutionary algorithm that incorporates both the idea
of simple, linear chromosomes of fixed length used in
Genetic Algorithms (GAs) and the structure of different
sizes and shapes used in Genetic Programming (GP). As
with other genetic programming algorithms, GEP has
difficulty finding appropriate numeric constants for
terminal nodes in the expression trees. In this paper,
we describe a new approach of constant generation using
Differential Evolution (DE), which is a simple
real-valued GA that has proven to be robust and
efficient on parameter optimisation problems. Our
experimental results on two symbolic regression
problems show that the approach significantly improves
the performance of the GEP algorithm. The proposed
approach can be easily extended to other Genetic
Programming variants.},
added-at = {2008-06-19T17:46:40.000+0200},
address = {Seattle, WA, USA},
author = {Zhang, Qiongyun and Zhou, Chi and Xiao, Weimin and Nelson, Peter C. and Li, Xin},
biburl = {https://www.bibsonomy.org/bibtex/20d01c5bfb43961426b94abbf3ffbf422/brazovayeye},
booktitle = {Late breaking paper at Genetic and Evolutionary
Computation Conference {(GECCO'2006)}},
editor = {Grahl, J{\"{o}}rn},
interhash = {c1b5c8ca3fa5daedc9c4a832563659de},
intrahash = {0d01c5bfb43961426b94abbf3ffbf422},
keywords = {DE algorithms, expression gene genetic programming,},
month = {8-12 July},
notes = {Distributed on CD-ROM at GECCO-2006},
timestamp = {2008-06-19T17:55:47.000+0200},
title = {Using Differential Evolution for {GEP} Constant
Creation},
url = {http://www.cs.bham.ac.uk/~wbl/biblio/gecco2006etc/papers/lbp130.pdf},
year = 2006
}