Publication detail

Indirectly Encoded Fitness Predictors Coevolved with Cartesian Programs

DRAHOŠOVÁ, M. HULVA, J. SEKANINA, L.

Original Title

Indirectly Encoded Fitness Predictors Coevolved with Cartesian Programs

English Title

Indirectly Encoded Fitness Predictors Coevolved with Cartesian Programs

Type

conference paper

Language

en

Original Abstract

We investigate coevolutionary Cartesian genetic programming that coevolves fitness predictors in order to diminish the number of target objective vector (TOV) evaluations, needed to obtain a satisfactory solution, to reduce the computational cost of evolution. This paper introduces the use of coevolution of fitness predictors in CGP with a new type of indirectly encoded predictors. Indirectly encoded predictors are operated using the CGP and provide a variable number of TOVs used for solution evaluation during the coevolution. It is shown in 5 symbolic regression problems that the proposed predictors are able to adapt the size of TOVs array in response to a particular training data set.

English abstract

We investigate coevolutionary Cartesian genetic programming that coevolves fitness predictors in order to diminish the number of target objective vector (TOV) evaluations, needed to obtain a satisfactory solution, to reduce the computational cost of evolution. This paper introduces the use of coevolution of fitness predictors in CGP with a new type of indirectly encoded predictors. Indirectly encoded predictors are operated using the CGP and provide a variable number of TOVs used for solution evaluation during the coevolution. It is shown in 5 symbolic regression problems that the proposed predictors are able to adapt the size of TOVs array in response to a particular training data set.

Keywords

coevolution, cartesian genetic programming, fitness prediction, symbolic regression

RIV year

2015

Released

15.03.2015

Publisher

Springer International Publishing

Location

Berlin

ISBN

978-3-319-16500-4

Book

Genetic Programming

Edition

Lecture Notes in Computer Science

Edition number

NEUVEDEN

Pages from

113

Pages to

125

Pages count

13

URL

Documents

BibTex


@inproceedings{BUT119803,
  author="Michaela {Drahošová} and Jiří {Hulva} and Lukáš {Sekanina}",
  title="Indirectly Encoded Fitness Predictors Coevolved with Cartesian Programs",
  annote="We investigate coevolutionary Cartesian genetic programming that coevolves
fitness predictors in order to diminish the number of target objective vector
(TOV) evaluations, needed to obtain a satisfactory solution, to reduce the
computational cost of evolution. This paper introduces the use of coevolution of
fitness predictors in CGP with a new type of indirectly encoded predictors.
Indirectly encoded predictors are operated using the CGP and provide a variable
number of TOVs used for solution evaluation during the coevolution. It is shown
in 5 symbolic regression problems that the proposed predictors are able to adapt
the size of TOVs array in response to a particular training data set.",
  address="Springer International Publishing",
  booktitle="Genetic Programming",
  chapter="119803",
  doi="10.1007/978-3-319-16501-1_10",
  edition="Lecture Notes in Computer Science",
  howpublished="print",
  institution="Springer International Publishing",
  year="2015",
  month="march",
  pages="113--125",
  publisher="Springer International Publishing",
  type="conference paper"
}