Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution

Bing Li, Xianneng Li, Shingo Mabu, Kotaro Hirasawa

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

Variable Size Genetic Network Programming (GNPvs) with Binomial Distribution is a new type of Genetic Network Programming, which permits the individuals with variable size and obtains the optimal size of individuals during evolution. In contrast to swapping the corresponding nodes in the crossover of Genetic Network Programming (GNP), the crossover of GNPvs moves a number of nodes from each parent individual to another parent individual. The probability of selecting the number of nodes to move satisfies the binomial distribution, which is controls by the crossover rate and the initial size of the individuals. In this paper, the performance of the algorithm and the size of the individuals are studied by changing the most important parameter, i.e., crossover rate. Since GNPvs may suffer from the bloating problem, the Island Model is introduced to control the size of the individuals implicitly. The well-known test bed Tileworld is used to show the numeric results in the simulations.

Original languageEnglish
Title of host publicationProceedings of the SICE Annual Conference
Pages155-160
Number of pages6
Publication statusPublished - 2011
Event50th Annual Conference on Society of Instrument and Control Engineers, SICE 2011 - Tokyo, Japan
Duration: 2011 Sep 132011 Sep 18

Other

Other50th Annual Conference on Society of Instrument and Control Engineers, SICE 2011
CountryJapan
CityTokyo
Period11/9/1311/9/18

Keywords

  • Binomial distribution
  • Bloating problem
  • Crossover
  • Genetic Network Programming
  • Variable size

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Control and Systems Engineering
  • Computer Science Applications

Cite this

Li, B., Li, X., Mabu, S., & Hirasawa, K. (2011). Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution. In Proceedings of the SICE Annual Conference (pp. 155-160). [6060594]

Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution. / Li, Bing; Li, Xianneng; Mabu, Shingo; Hirasawa, Kotaro.

Proceedings of the SICE Annual Conference. 2011. p. 155-160 6060594.

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Li, B, Li, X, Mabu, S & Hirasawa, K 2011, Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution. in Proceedings of the SICE Annual Conference., 6060594, pp. 155-160, 50th Annual Conference on Society of Instrument and Control Engineers, SICE 2011, Tokyo, Japan, 11/9/13.
Li B, Li X, Mabu S, Hirasawa K. Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution. In Proceedings of the SICE Annual Conference. 2011. p. 155-160. 6060594
Li, Bing ; Li, Xianneng ; Mabu, Shingo ; Hirasawa, Kotaro. / Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution. Proceedings of the SICE Annual Conference. 2011. pp. 155-160
@inproceedings{0aa25f78b58d40d8925fa53bfa6bfe07,
title = "Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution",
abstract = "Variable Size Genetic Network Programming (GNPvs) with Binomial Distribution is a new type of Genetic Network Programming, which permits the individuals with variable size and obtains the optimal size of individuals during evolution. In contrast to swapping the corresponding nodes in the crossover of Genetic Network Programming (GNP), the crossover of GNPvs moves a number of nodes from each parent individual to another parent individual. The probability of selecting the number of nodes to move satisfies the binomial distribution, which is controls by the crossover rate and the initial size of the individuals. In this paper, the performance of the algorithm and the size of the individuals are studied by changing the most important parameter, i.e., crossover rate. Since GNPvs may suffer from the bloating problem, the Island Model is introduced to control the size of the individuals implicitly. The well-known test bed Tileworld is used to show the numeric results in the simulations.",
keywords = "Binomial distribution, Bloating problem, Crossover, Genetic Network Programming, Variable size",
author = "Bing Li and Xianneng Li and Shingo Mabu and Kotaro Hirasawa",
year = "2011",
language = "English",
isbn = "9784907764395",
pages = "155--160",
booktitle = "Proceedings of the SICE Annual Conference",

}

TY - GEN

T1 - Analysis of crossover rate in variable size Genetic Network Programming with binomial distribution

AU - Li, Bing

AU - Li, Xianneng

AU - Mabu, Shingo

AU - Hirasawa, Kotaro

PY - 2011

Y1 - 2011

N2 - Variable Size Genetic Network Programming (GNPvs) with Binomial Distribution is a new type of Genetic Network Programming, which permits the individuals with variable size and obtains the optimal size of individuals during evolution. In contrast to swapping the corresponding nodes in the crossover of Genetic Network Programming (GNP), the crossover of GNPvs moves a number of nodes from each parent individual to another parent individual. The probability of selecting the number of nodes to move satisfies the binomial distribution, which is controls by the crossover rate and the initial size of the individuals. In this paper, the performance of the algorithm and the size of the individuals are studied by changing the most important parameter, i.e., crossover rate. Since GNPvs may suffer from the bloating problem, the Island Model is introduced to control the size of the individuals implicitly. The well-known test bed Tileworld is used to show the numeric results in the simulations.

AB - Variable Size Genetic Network Programming (GNPvs) with Binomial Distribution is a new type of Genetic Network Programming, which permits the individuals with variable size and obtains the optimal size of individuals during evolution. In contrast to swapping the corresponding nodes in the crossover of Genetic Network Programming (GNP), the crossover of GNPvs moves a number of nodes from each parent individual to another parent individual. The probability of selecting the number of nodes to move satisfies the binomial distribution, which is controls by the crossover rate and the initial size of the individuals. In this paper, the performance of the algorithm and the size of the individuals are studied by changing the most important parameter, i.e., crossover rate. Since GNPvs may suffer from the bloating problem, the Island Model is introduced to control the size of the individuals implicitly. The well-known test bed Tileworld is used to show the numeric results in the simulations.

KW - Binomial distribution

KW - Bloating problem

KW - Crossover

KW - Genetic Network Programming

KW - Variable size

UR - http://www.scopus.com/inward/record.url?scp=81255179820&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=81255179820&partnerID=8YFLogxK

M3 - Conference contribution

SN - 9784907764395

SP - 155

EP - 160

BT - Proceedings of the SICE Annual Conference

ER -