Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization

Selection HyperHeuristics are informed search methods that work in a higher abstraction level than heuristic or MetaHeuristics: they constitute heuristics to choose heuristics. Such selection is realized by a Choice Function (CF), whose target is to decide which heuristic strategy is applied in each...

Descripción completa

Detalles Bibliográficos
Autores principales: Casanova, Carlos, Schweickardt, Gustavo, Camargo, Federico
Formato: Artículo revista
Lenguaje:Español
Publicado: Escuela de Perfeccionamiento en Investigación Operativa 2018
Materias:
artificial neural networks
case based reasoning
particle swarm optimization
selection hyperheuristics
soft computing
redes neuronales artificiales
razonamiento basado en casos
optimización por enjambre de partículas
hiperheurísticas de selección
Acceso en línea:https://revistas.unc.edu.ar/index.php/epio/article/view/22200
Aporte de:
Revista de la Escuela de Perfeccionamiento en Investigación Operativa de Universidad Nacional de Córdoba
id I10-R359-article-22200
record_format ojs
spelling I10-R359-article-222002018-12-03T15:47:55Z Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization Modelo hiperheurístico HY X-FPSO CBR SII soportado en metaheurísticas X-PSO multiobjetivo para resolver una clase de problemas de optimización combinatoria Casanova, Carlos Schweickardt, Gustavo Camargo, Federico artificial neural networks case based reasoning particle swarm optimization selection hyperheuristics soft computing redes neuronales artificiales razonamiento basado en casos optimización por enjambre de partículas hiperheurísticas de selección soft computing Selection HyperHeuristics are informed search methods that work in a higher abstraction level than heuristic or MetaHeuristics: they constitute heuristics to choose heuristics. Such selection is realized by a Choice Function (CF), whose target is to decide which heuristic strategy is applied in each decision instance of the algorithm, using for that non-domain data about the problem being solved. In this work a Case Based Reasoning Selection HyperHeuristic with X-PSO MultiObjective domain is presented, whose CF is constituted of a Feed-Forward Artificial Neural Network (ANN) of Multi-Layer Perceptron (MLP) type. The non-domain information used by the CF is composed of Swarm Intelligence Indicators, proposed by the authors in previous papers, which aims to give a measure on the abilities of a swarm to solve a particular problem. The design and the optimization problem associated to the CF Case Based Training are presented, so as the method to carry out such training. Finally, the process is applied to build a CF for a CBR Hyperheuristic that solves two Combinatorial Optimization Problems: the Load Balancing of a Three Phase Power Distribution System and the Reliability Optimization of Electrical Distribution Systems in Medium-Voltage. Las Hiperheurísticas de Selección constituyen métodos de búsqueda concebidos en un nivel de abstracción superior al de las MetaHeurísticas. Para ello, una Función de Selección (FS), cuyo objetivo es decidir cuál de las estrategias MetaHeurísticas se aplica en cada instancia de decisión, evalúa la  aptitud de las mismas en cada solución iterativa. En este trabajo se presenta una HiperHeurística de Selección Basada en Razonamiento (CBR) con dominio en MetaHeurísticas X-PSO MultiObjetivo, HY X-FPSO CBR SII, cuya FS se constituye de una Red Neuronal Artificial (RN) de propagación hacia adelante tipo Multi-Layer Perceptron (MLP). La información utilizada por la FS proviene desde  Indicadores de Inteligencia de Grupo, propuestos por los autores en trabajos previos, que proporcionan una medida de la habilidad de cada MetaHeurística para resolver cierta instancia del problema. Se aborda el diseño de la FS y el método de optimización asociado al Entrenamiento Basado en Casos de la misma. Este novedoso enfoque, aporte principal del trabajo,  permite construir una única FS capaz de resolver dos problemas de optimización combinatoria: el Balance de Cargas de un Sistema Trifásico de Distribución de Energía Eléctrica (SDEE) y la Optimización de la Confiabilidad de un SDEE en Media Tensión. Escuela de Perfeccionamiento en Investigación Operativa 2018-11-30 info:eu-repo/semantics/article info:eu-repo/semantics/publishedVersion application/pdf https://revistas.unc.edu.ar/index.php/epio/article/view/22200 Revista de la Escuela de Perfeccionamiento en Investigación Operativa; Vol. 26 Núm. 44 (2018): Noviembre; 4-20 1853-9777 0329-7322 spa https://revistas.unc.edu.ar/index.php/epio/article/view/22200/21806
institution Universidad Nacional de Córdoba
institution_str I-10
repository_str R-359
container_title_str Revista de la Escuela de Perfeccionamiento en Investigación Operativa
language Español
format Artículo revista
topic artificial neural networks
case based reasoning
particle swarm optimization
selection hyperheuristics
soft computing
redes neuronales artificiales
razonamiento basado en casos
optimización por enjambre de partículas
hiperheurísticas de selección
soft computing
spellingShingle artificial neural networks
case based reasoning
particle swarm optimization
selection hyperheuristics
soft computing
redes neuronales artificiales
razonamiento basado en casos
optimización por enjambre de partículas
hiperheurísticas de selección
soft computing
Casanova, Carlos
Schweickardt, Gustavo
Camargo, Federico
Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization
topic_facet artificial neural networks
case based reasoning
particle swarm optimization
selection hyperheuristics
soft computing
redes neuronales artificiales
razonamiento basado en casos
optimización por enjambre de partículas
hiperheurísticas de selección
soft computing
author Casanova, Carlos
Schweickardt, Gustavo
Camargo, Federico
author_facet Casanova, Carlos
Schweickardt, Gustavo
Camargo, Federico
author_sort Casanova, Carlos
title Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization
title_short Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization
title_full Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization
title_fullStr Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization
title_full_unstemmed Hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization
title_sort hyperheuristic model hy x-fpso cbr sii supported in metaheuristics x-pso multiobjective to solve a class of problems of combinatorial optimization
description Selection HyperHeuristics are informed search methods that work in a higher abstraction level than heuristic or MetaHeuristics: they constitute heuristics to choose heuristics. Such selection is realized by a Choice Function (CF), whose target is to decide which heuristic strategy is applied in each decision instance of the algorithm, using for that non-domain data about the problem being solved. In this work a Case Based Reasoning Selection HyperHeuristic with X-PSO MultiObjective domain is presented, whose CF is constituted of a Feed-Forward Artificial Neural Network (ANN) of Multi-Layer Perceptron (MLP) type. The non-domain information used by the CF is composed of Swarm Intelligence Indicators, proposed by the authors in previous papers, which aims to give a measure on the abilities of a swarm to solve a particular problem. The design and the optimization problem associated to the CF Case Based Training are presented, so as the method to carry out such training. Finally, the process is applied to build a CF for a CBR Hyperheuristic that solves two Combinatorial Optimization Problems: the Load Balancing of a Three Phase Power Distribution System and the Reliability Optimization of Electrical Distribution Systems in Medium-Voltage.
publisher Escuela de Perfeccionamiento en Investigación Operativa
publishDate 2018
url https://revistas.unc.edu.ar/index.php/epio/article/view/22200
work_keys_str_mv AT casanovacarlos hyperheuristicmodelhyxfpsocbrsiisupportedinmetaheuristicsxpsomultiobjectivetosolveaclassofproblemsofcombinatorialoptimization
AT schweickardtgustavo hyperheuristicmodelhyxfpsocbrsiisupportedinmetaheuristicsxpsomultiobjectivetosolveaclassofproblemsofcombinatorialoptimization
AT camargofederico hyperheuristicmodelhyxfpsocbrsiisupportedinmetaheuristicsxpsomultiobjectivetosolveaclassofproblemsofcombinatorialoptimization
AT casanovacarlos modelohiperheuristicohyxfpsocbrsiisoportadoenmetaheuristicasxpsomultiobjetivopararesolverunaclasedeproblemasdeoptimizacioncombinatoria
AT schweickardtgustavo modelohiperheuristicohyxfpsocbrsiisoportadoenmetaheuristicasxpsomultiobjetivopararesolverunaclasedeproblemasdeoptimizacioncombinatoria
AT camargofederico modelohiperheuristicohyxfpsocbrsiisoportadoenmetaheuristicasxpsomultiobjetivopararesolverunaclasedeproblemasdeoptimizacioncombinatoria
first_indexed 2024-09-03T22:23:23Z
last_indexed 2024-09-03T22:23:23Z
_version_ 1809215340879544320

Ejemplares similares