Tuners review

How crucial are set-up values to find effective parameter values?

Elizabeth Montero, María Cristina Riff, Nicolás Rojas-Morales

Resultado de la investigación: Article

2 Citas (Scopus)

Resumen

ParamILS, I-Race and Evoca are well-known tuning methods designed to search quality parameter calibrations for metaheuristic algorithms. The set-up of parameter search space can strongly affect the performance of tuning methods. In this work we study how the parameter search definitions affect the quality of parameter calibrations delivered by these tuners. An experimental evaluation using two well known metaheuristic algorithms and a real life case is presented. We also provide some guidelines to consider when defining parameters search spaces according to the tuner used in order to obtain the best performance they can find.

Idioma originalEnglish
Páginas (desde-hasta)108-118
Número de páginas11
PublicaciónEngineering Applications of Artificial Intelligence
Volumen76
DOI
EstadoPublished - 1 nov 2018

Huella dactilar

Tuners
Tuning
Calibration

ASJC Scopus subject areas

  • Control and Systems Engineering
  • Artificial Intelligence
  • Electrical and Electronic Engineering

Citar esto

@article{6183bb6cd9a14769a27d5d77fe8507a4,
title = "Tuners review: How crucial are set-up values to find effective parameter values?",
abstract = "ParamILS, I-Race and Evoca are well-known tuning methods designed to search quality parameter calibrations for metaheuristic algorithms. The set-up of parameter search space can strongly affect the performance of tuning methods. In this work we study how the parameter search definitions affect the quality of parameter calibrations delivered by these tuners. An experimental evaluation using two well known metaheuristic algorithms and a real life case is presented. We also provide some guidelines to consider when defining parameters search spaces according to the tuner used in order to obtain the best performance they can find.",
keywords = "Parameter search space definition, Parameter setting, Tuning methods",
author = "Elizabeth Montero and Riff, {Mar{\'i}a Cristina} and Nicol{\'a}s Rojas-Morales",
year = "2018",
month = "11",
day = "1",
doi = "10.1016/j.engappai.2018.09.001",
language = "English",
volume = "76",
pages = "108--118",
journal = "Engineering Applications of Artificial Intelligence",
issn = "0952-1976",
publisher = "Elsevier Limited",

}

Tuners review : How crucial are set-up values to find effective parameter values? / Montero, Elizabeth; Riff, María Cristina; Rojas-Morales, Nicolás.

En: Engineering Applications of Artificial Intelligence, Vol. 76, 01.11.2018, p. 108-118.

Resultado de la investigación: Article

TY - JOUR

T1 - Tuners review

T2 - How crucial are set-up values to find effective parameter values?

AU - Montero, Elizabeth

AU - Riff, María Cristina

AU - Rojas-Morales, Nicolás

PY - 2018/11/1

Y1 - 2018/11/1

N2 - ParamILS, I-Race and Evoca are well-known tuning methods designed to search quality parameter calibrations for metaheuristic algorithms. The set-up of parameter search space can strongly affect the performance of tuning methods. In this work we study how the parameter search definitions affect the quality of parameter calibrations delivered by these tuners. An experimental evaluation using two well known metaheuristic algorithms and a real life case is presented. We also provide some guidelines to consider when defining parameters search spaces according to the tuner used in order to obtain the best performance they can find.

AB - ParamILS, I-Race and Evoca are well-known tuning methods designed to search quality parameter calibrations for metaheuristic algorithms. The set-up of parameter search space can strongly affect the performance of tuning methods. In this work we study how the parameter search definitions affect the quality of parameter calibrations delivered by these tuners. An experimental evaluation using two well known metaheuristic algorithms and a real life case is presented. We also provide some guidelines to consider when defining parameters search spaces according to the tuner used in order to obtain the best performance they can find.

KW - Parameter search space definition

KW - Parameter setting

KW - Tuning methods

UR - http://www.scopus.com/inward/record.url?scp=85053758543&partnerID=8YFLogxK

U2 - 10.1016/j.engappai.2018.09.001

DO - 10.1016/j.engappai.2018.09.001

M3 - Article

VL - 76

SP - 108

EP - 118

JO - Engineering Applications of Artificial Intelligence

JF - Engineering Applications of Artificial Intelligence

SN - 0952-1976

ER -