A Study of Good Predecessor Programs for Reducing
Fitness Evaluation Cost in Genetic Programming
H. Xie, M. Zhang, and P. Andreae. CS-TR-06-3. Computer Science, Victoria University of Wellington, New Zealand, (2006)
Abstract
Good Predecessor Programs (GPPs) are the ancestors of
the best program found in a Genetic Programming (GP)
evolution. This paper reports on an investigation into
GPPs with the ultimate goal of reducing fitness
evaluation cost in tree-based GP systems. A framework
is developed for gathering information about GPPs and a
series of experiments is conducted on a symbolic
regression problem, a binary classification problem,
and a multi-class classification program with
increasing levels of difficulty in different domains.
The analysis of the data shows that during evolution,
GPPs typically constitute between less than 33per cent
of the total programs evaluated, and may constitute
less than 5per cent. The analysis results further shows
that in all evaluated programs, the proportion of GPPs
is reduced by increasing tournament size and to a less
extent, affected by population size. Problem difficulty
seems to have no clear influence on the proportion of
GPPs.
%0 Report
%1 vuw-CS-TR-06-3
%A Xie, Huayang
%A Zhang, Mengjie
%A Andreae, Peter
%C New Zealand
%D 2006
%K Fitness algorithms, clustering evaluation, genetic good population predecessor programming, programs,
%N CS-TR-06-3
%T A Study of Good Predecessor Programs for Reducing
Fitness Evaluation Cost in Genetic Programming
%U http://www.mcs.vuw.ac.nz/comp/Publications/CS-TR-06-3.abs.html
%X Good Predecessor Programs (GPPs) are the ancestors of
the best program found in a Genetic Programming (GP)
evolution. This paper reports on an investigation into
GPPs with the ultimate goal of reducing fitness
evaluation cost in tree-based GP systems. A framework
is developed for gathering information about GPPs and a
series of experiments is conducted on a symbolic
regression problem, a binary classification problem,
and a multi-class classification program with
increasing levels of difficulty in different domains.
The analysis of the data shows that during evolution,
GPPs typically constitute between less than 33per cent
of the total programs evaluated, and may constitute
less than 5per cent. The analysis results further shows
that in all evaluated programs, the proportion of GPPs
is reduced by increasing tournament size and to a less
extent, affected by population size. Problem difficulty
seems to have no clear influence on the proportion of
GPPs.
@techreport{vuw-CS-TR-06-3,
abstract = {Good Predecessor Programs (GPPs) are the ancestors of
the best program found in a Genetic Programming (GP)
evolution. This paper reports on an investigation into
GPPs with the ultimate goal of reducing fitness
evaluation cost in tree-based GP systems. A framework
is developed for gathering information about GPPs and a
series of experiments is conducted on a symbolic
regression problem, a binary classification problem,
and a multi-class classification program with
increasing levels of difficulty in different domains.
The analysis of the data shows that during evolution,
GPPs typically constitute between less than 33per cent
of the total programs evaluated, and may constitute
less than 5per cent. The analysis results further shows
that in all evaluated programs, the proportion of GPPs
is reduced by increasing tournament size and to a less
extent, affected by population size. Problem difficulty
seems to have no clear influence on the proportion of
GPPs.},
added-at = {2008-06-19T17:35:00.000+0200},
address = {New Zealand},
author = {Xie, Huayang and Zhang, Mengjie and Andreae, Peter},
biburl = {https://www.bibsonomy.org/bibtex/24f24dbcf0e5441a11037fb4ee631888f/brazovayeye},
institution = {Computer Science, Victoria University of Wellington},
interhash = {ccb4b966956a90ddc4b525c63b7636bd},
intrahash = {4f24dbcf0e5441a11037fb4ee631888f},
keywords = {Fitness algorithms, clustering evaluation, genetic good population predecessor programming, programs,},
number = {CS-TR-06-3},
timestamp = {2008-06-19T17:54:35.000+0200},
title = {A Study of Good Predecessor Programs for Reducing
Fitness Evaluation Cost in Genetic Programming},
url = {http://www.mcs.vuw.ac.nz/comp/Publications/CS-TR-06-3.abs.html},
year = 2006
}