An hybrid neural/genetic approach to continuous multi-objective optimization problems

Mario Costa, Edmondo Minisci, Eros Pasero

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

3 Citations (Scopus)

Abstract

Evolutionary algorithms perform optimization using the information derived from a population of sample solution points. Recent developments in this field regard optimization as the evolutionary process of an explicit, probabilistic model of the search space. The algorithms derived on the basis of this new philosophy maintain every feature of the classic evolutionary algorithms, but arc able to overcome some drawbacks. In this paper an evolutionary multi-objective optimization tool based on an estimation of distribution algorithm is proposed. It uses the ranking method of non-dominated sorting genetic algorithm-II and the Parzen estimator to approximate the probability density of solutions lying on the Pareto front. The proposed algorithm has been applied to different types of test case problems and results show good performance of the overall optimization procedure in terms of the number of function evaluations.

LanguageEnglish
Title of host publicationNeural Nets
Subtitle of host publication14th Italian Workshop on Neural Nets, WIRN VIETRI 2003, Vietri sul Mare, Italy, June 4-7, 2003. Revised Papers
EditorsBruno Apolloni, Maria Marinaro, Roberto Tagliaferri
Place of PublicationBerlin
PublisherSpringer
Pages61-69
Number of pages9
ISBN (Print)9783540202271, 9783540452164
DOIs
Publication statusPublished - 7 Jun 2003
Event14th Italian Workshop on Neural Nets - , Italy
Duration: 4 Jun 20037 Jun 2003

Publication series

NameLecture Notes in Computer Science
PublisherSpringer Verlag
Volume2859
ISSN (Print)0302-9743

Conference

Conference14th Italian Workshop on Neural Nets
CountryItaly
Period4/06/037/06/03

Fingerprint

Continuous Optimization
Multiobjective Optimization Problems
Multiobjective optimization
Evolutionary algorithms
Optimization
Evolutionary Algorithms
Evolutionary multiobjective Optimization
Function evaluation
Sorting algorithm
Pareto Front
Evaluation Function
Probability Density
Sorting
Probabilistic Model
Search Space
Ranking
Arc of a curve
Genetic algorithms
Genetic Algorithm
Estimator

Keywords

  • probability density function
  • Pareto front
  • distribution algorithm
  • objective function evaluation
  • Bayesian optimization algorithm

Cite this

Costa, M., Minisci, E., & Pasero, E. (2003). An hybrid neural/genetic approach to continuous multi-objective optimization problems. In B. Apolloni, M. Marinaro, & R. Tagliaferri (Eds.), Neural Nets: 14th Italian Workshop on Neural Nets, WIRN VIETRI 2003, Vietri sul Mare, Italy, June 4-7, 2003. Revised Papers (pp. 61-69). (Lecture Notes in Computer Science; Vol. 2859). Berlin: Springer. https://doi.org/10.1007/978-3-540-45216-4_6
Costa, Mario ; Minisci, Edmondo ; Pasero, Eros. / An hybrid neural/genetic approach to continuous multi-objective optimization problems. Neural Nets: 14th Italian Workshop on Neural Nets, WIRN VIETRI 2003, Vietri sul Mare, Italy, June 4-7, 2003. Revised Papers. editor / Bruno Apolloni ; Maria Marinaro ; Roberto Tagliaferri. Berlin : Springer, 2003. pp. 61-69 (Lecture Notes in Computer Science).
@inproceedings{e6a35791a54641678cdeaa5941e0e708,
title = "An hybrid neural/genetic approach to continuous multi-objective optimization problems",
abstract = "Evolutionary algorithms perform optimization using the information derived from a population of sample solution points. Recent developments in this field regard optimization as the evolutionary process of an explicit, probabilistic model of the search space. The algorithms derived on the basis of this new philosophy maintain every feature of the classic evolutionary algorithms, but arc able to overcome some drawbacks. In this paper an evolutionary multi-objective optimization tool based on an estimation of distribution algorithm is proposed. It uses the ranking method of non-dominated sorting genetic algorithm-II and the Parzen estimator to approximate the probability density of solutions lying on the Pareto front. The proposed algorithm has been applied to different types of test case problems and results show good performance of the overall optimization procedure in terms of the number of function evaluations.",
keywords = "probability density function, Pareto front, distribution algorithm, objective function evaluation, Bayesian optimization algorithm",
author = "Mario Costa and Edmondo Minisci and Eros Pasero",
year = "2003",
month = "6",
day = "7",
doi = "10.1007/978-3-540-45216-4_6",
language = "English",
isbn = "9783540202271",
series = "Lecture Notes in Computer Science",
publisher = "Springer",
pages = "61--69",
editor = "Bruno Apolloni and Maria Marinaro and Roberto Tagliaferri",
booktitle = "Neural Nets",

}

Costa, M, Minisci, E & Pasero, E 2003, An hybrid neural/genetic approach to continuous multi-objective optimization problems. in B Apolloni, M Marinaro & R Tagliaferri (eds), Neural Nets: 14th Italian Workshop on Neural Nets, WIRN VIETRI 2003, Vietri sul Mare, Italy, June 4-7, 2003. Revised Papers. Lecture Notes in Computer Science, vol. 2859, Springer, Berlin, pp. 61-69, 14th Italian Workshop on Neural Nets, Italy, 4/06/03. https://doi.org/10.1007/978-3-540-45216-4_6

An hybrid neural/genetic approach to continuous multi-objective optimization problems. / Costa, Mario; Minisci, Edmondo; Pasero, Eros.

Neural Nets: 14th Italian Workshop on Neural Nets, WIRN VIETRI 2003, Vietri sul Mare, Italy, June 4-7, 2003. Revised Papers. ed. / Bruno Apolloni; Maria Marinaro; Roberto Tagliaferri. Berlin : Springer, 2003. p. 61-69 (Lecture Notes in Computer Science; Vol. 2859).

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

TY - GEN

T1 - An hybrid neural/genetic approach to continuous multi-objective optimization problems

AU - Costa, Mario

AU - Minisci, Edmondo

AU - Pasero, Eros

PY - 2003/6/7

Y1 - 2003/6/7

N2 - Evolutionary algorithms perform optimization using the information derived from a population of sample solution points. Recent developments in this field regard optimization as the evolutionary process of an explicit, probabilistic model of the search space. The algorithms derived on the basis of this new philosophy maintain every feature of the classic evolutionary algorithms, but arc able to overcome some drawbacks. In this paper an evolutionary multi-objective optimization tool based on an estimation of distribution algorithm is proposed. It uses the ranking method of non-dominated sorting genetic algorithm-II and the Parzen estimator to approximate the probability density of solutions lying on the Pareto front. The proposed algorithm has been applied to different types of test case problems and results show good performance of the overall optimization procedure in terms of the number of function evaluations.

AB - Evolutionary algorithms perform optimization using the information derived from a population of sample solution points. Recent developments in this field regard optimization as the evolutionary process of an explicit, probabilistic model of the search space. The algorithms derived on the basis of this new philosophy maintain every feature of the classic evolutionary algorithms, but arc able to overcome some drawbacks. In this paper an evolutionary multi-objective optimization tool based on an estimation of distribution algorithm is proposed. It uses the ranking method of non-dominated sorting genetic algorithm-II and the Parzen estimator to approximate the probability density of solutions lying on the Pareto front. The proposed algorithm has been applied to different types of test case problems and results show good performance of the overall optimization procedure in terms of the number of function evaluations.

KW - probability density function

KW - Pareto front

KW - distribution algorithm

KW - objective function evaluation

KW - Bayesian optimization algorithm

UR - http://www.scopus.com/inward/record.url?scp=0142218850&partnerID=8YFLogxK

U2 - 10.1007/978-3-540-45216-4_6

DO - 10.1007/978-3-540-45216-4_6

M3 - Conference contribution book

SN - 9783540202271

SN - 9783540452164

T3 - Lecture Notes in Computer Science

SP - 61

EP - 69

BT - Neural Nets

A2 - Apolloni, Bruno

A2 - Marinaro, Maria

A2 - Tagliaferri, Roberto

PB - Springer

CY - Berlin

ER -

Costa M, Minisci E, Pasero E. An hybrid neural/genetic approach to continuous multi-objective optimization problems. In Apolloni B, Marinaro M, Tagliaferri R, editors, Neural Nets: 14th Italian Workshop on Neural Nets, WIRN VIETRI 2003, Vietri sul Mare, Italy, June 4-7, 2003. Revised Papers. Berlin: Springer. 2003. p. 61-69. (Lecture Notes in Computer Science). https://doi.org/10.1007/978-3-540-45216-4_6