Trading accuracy for simplicity in evolutionary algorithms

Maria Cristina Riff, Elizabeth Montero

Resultado de la investigación: Conference contribution

Resumen

Many evolutionary algorithms efficient to solve a wide range of problems have been proposed and validated in the literature. We call the problem addressed in this paper 'the problem of refining the design'. Given an evolutionary algorithm that includes many operators, we would like to assess if all of the latter are really relevant to accomplish the former's performance. In this paper, we present a framework to evaluate an evolutionary algorithm that has already been designed. The goal of this evaluation is to study the choices that we have to simplify the code in respect to its results accuracy. The results show that an efficient tuner can assist us in this task by obtaining information that help design decisions.

Idioma originalEnglish
Título de la publicación alojada2013 IEEE Congress on Evolutionary Computation, CEC 2013
Páginas3125-3131
Número de páginas7
DOI
EstadoPublished - 21 ago 2013
Evento2013 IEEE Congress on Evolutionary Computation, CEC 2013 - Cancun, Mexico
Duración: 20 jun 201323 jun 2013

Conference

Conference2013 IEEE Congress on Evolutionary Computation, CEC 2013
PaísMexico
CiudadCancun
Período20/06/1323/06/13

Huella dactilar

Evolutionary algorithms
Evolutionary Algorithms
Simplicity
Refining
Mathematical operators
Simplify
Evaluate
Evaluation
Operator
Range of data
Design
Framework

ASJC Scopus subject areas

  • Computational Theory and Mathematics
  • Theoretical Computer Science

Citar esto

Riff, M. C., & Montero, E. (2013). Trading accuracy for simplicity in evolutionary algorithms. En 2013 IEEE Congress on Evolutionary Computation, CEC 2013 (pp. 3125-3131). [6557951] https://doi.org/10.1109/CEC.2013.6557951
Riff, Maria Cristina ; Montero, Elizabeth. / Trading accuracy for simplicity in evolutionary algorithms. 2013 IEEE Congress on Evolutionary Computation, CEC 2013. 2013. pp. 3125-3131
@inproceedings{e5f386914e8d422cab7e7b98afc67fce,
title = "Trading accuracy for simplicity in evolutionary algorithms",
abstract = "Many evolutionary algorithms efficient to solve a wide range of problems have been proposed and validated in the literature. We call the problem addressed in this paper 'the problem of refining the design'. Given an evolutionary algorithm that includes many operators, we would like to assess if all of the latter are really relevant to accomplish the former's performance. In this paper, we present a framework to evaluate an evolutionary algorithm that has already been designed. The goal of this evaluation is to study the choices that we have to simplify the code in respect to its results accuracy. The results show that an efficient tuner can assist us in this task by obtaining information that help design decisions.",
author = "Riff, {Maria Cristina} and Elizabeth Montero",
year = "2013",
month = "8",
day = "21",
doi = "10.1109/CEC.2013.6557951",
language = "English",
isbn = "9781479904549",
pages = "3125--3131",
booktitle = "2013 IEEE Congress on Evolutionary Computation, CEC 2013",

}

Riff, MC & Montero, E 2013, Trading accuracy for simplicity in evolutionary algorithms. En 2013 IEEE Congress on Evolutionary Computation, CEC 2013., 6557951, pp. 3125-3131, 2013 IEEE Congress on Evolutionary Computation, CEC 2013, Cancun, Mexico, 20/06/13. https://doi.org/10.1109/CEC.2013.6557951

Trading accuracy for simplicity in evolutionary algorithms. / Riff, Maria Cristina; Montero, Elizabeth.

2013 IEEE Congress on Evolutionary Computation, CEC 2013. 2013. p. 3125-3131 6557951.

Resultado de la investigación: Conference contribution

TY - GEN

T1 - Trading accuracy for simplicity in evolutionary algorithms

AU - Riff, Maria Cristina

AU - Montero, Elizabeth

PY - 2013/8/21

Y1 - 2013/8/21

N2 - Many evolutionary algorithms efficient to solve a wide range of problems have been proposed and validated in the literature. We call the problem addressed in this paper 'the problem of refining the design'. Given an evolutionary algorithm that includes many operators, we would like to assess if all of the latter are really relevant to accomplish the former's performance. In this paper, we present a framework to evaluate an evolutionary algorithm that has already been designed. The goal of this evaluation is to study the choices that we have to simplify the code in respect to its results accuracy. The results show that an efficient tuner can assist us in this task by obtaining information that help design decisions.

AB - Many evolutionary algorithms efficient to solve a wide range of problems have been proposed and validated in the literature. We call the problem addressed in this paper 'the problem of refining the design'. Given an evolutionary algorithm that includes many operators, we would like to assess if all of the latter are really relevant to accomplish the former's performance. In this paper, we present a framework to evaluate an evolutionary algorithm that has already been designed. The goal of this evaluation is to study the choices that we have to simplify the code in respect to its results accuracy. The results show that an efficient tuner can assist us in this task by obtaining information that help design decisions.

UR - http://www.scopus.com/inward/record.url?scp=84881567957&partnerID=8YFLogxK

U2 - 10.1109/CEC.2013.6557951

DO - 10.1109/CEC.2013.6557951

M3 - Conference contribution

AN - SCOPUS:84881567957

SN - 9781479904549

SP - 3125

EP - 3131

BT - 2013 IEEE Congress on Evolutionary Computation, CEC 2013

ER -

Riff MC, Montero E. Trading accuracy for simplicity in evolutionary algorithms. En 2013 IEEE Congress on Evolutionary Computation, CEC 2013. 2013. p. 3125-3131. 6557951 https://doi.org/10.1109/CEC.2013.6557951