In this paper a method to increase the optimization ability of genetic algorithms (GAs) is proposed. To promote population diversity, a fraction of the worst individuals of the current population is replaced by individuals from an older population. To experimentally validate the approach we have used a set of well-known benchmark problems of tunable difficulty for GAs. Standard GA with and without elitism and steady state GA have been augmented with the proposed method. The obtained results show that the algorithms augmented with the proposed method perform better than the not-augmented algorithms or have the same performances. Furthermore, the proposed method depends on two parameters: one of them regulates the size of the fraction of the population replaced and the other one decides the "age" of the population used for the replacement. Experimental results indicate that better performances have been achieved with high values of the former parameter and low values of the latter one. © 2011 Springer-Verlag

Castelli, M., Manzoni, L., Vanneschi, L. (2011). A method to reuse old populations in genetic algorithms. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp.138-152). SPRINGER-VERLAG BERLIN [10.1007/978-3-642-24769-9_11].

A method to reuse old populations in genetic algorithms

CASTELLI, MAURO;MANZONI, LUCA;VANNESCHI, LEONARDO
2011

Abstract

In this paper a method to increase the optimization ability of genetic algorithms (GAs) is proposed. To promote population diversity, a fraction of the worst individuals of the current population is replaced by individuals from an older population. To experimentally validate the approach we have used a set of well-known benchmark problems of tunable difficulty for GAs. Standard GA with and without elitism and steady state GA have been augmented with the proposed method. The obtained results show that the algorithms augmented with the proposed method perform better than the not-augmented algorithms or have the same performances. Furthermore, the proposed method depends on two parameters: one of them regulates the size of the fraction of the population replaced and the other one decides the "age" of the population used for the replacement. Experimental results indicate that better performances have been achieved with high values of the former parameter and low values of the latter one. © 2011 Springer-Verlag
No
paper
Computer Science (all); Theoretical Computer Science
English
15th Portuguese Conference on Artificial Intelligence, EPIA 2011
9783642247682
Castelli, M., Manzoni, L., Vanneschi, L. (2011). A method to reuse old populations in genetic algorithms. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (pp.138-152). SPRINGER-VERLAG BERLIN [10.1007/978-3-642-24769-9_11].
Castelli, M; Manzoni, L; Vanneschi, L
File in questo prodotto:
Non ci sono file associati a questo prodotto.

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/10281/60794
Citazioni
  • Scopus 2
  • ???jsp.display-item.citation.isi??? 2
Social impact