Using a well known benchmarking and profiling environment, we compare the performances of three simple and easy to use metaheuristics for global optimization: Differential Evolution, Basin Hopping and Particle Swarm Optimization. The comparison was done on a test set of 24 functions featuring many characteristics found on real-world problems and on four different space dimensions. Our results statistically show that there is no clear winner overall. The three methods perform well in general and the actual differences are related to the different groups of functions in the benchmark with Basin Hopping being the most robust technique, and Differential Evolution and Particle Swarm Optimization excelling on highly multi-modal functions.

Comparing Basin Hopping with Differential Evolution and Particle Swarm Optimization

Santucci, Valentino
;
2022

Abstract

Using a well known benchmarking and profiling environment, we compare the performances of three simple and easy to use metaheuristics for global optimization: Differential Evolution, Basin Hopping and Particle Swarm Optimization. The comparison was done on a test set of 24 functions featuring many characteristics found on real-world problems and on four different space dimensions. Our results statistically show that there is no clear winner overall. The three methods perform well in general and the actual differences are related to the different groups of functions in the benchmark with Basin Hopping being the most robust technique, and Differential Evolution and Particle Swarm Optimization excelling on highly multi-modal functions.
978-3-031-02461-0
978-3-031-02462-7
Global continuous optimization, Metaheuristics, Benchmarking
File in questo prodotto:
File Dimensione Formato  
EvoStar_2022___Basin_Hopping.pdf

embargo fino al 15/04/2023

Descrizione: Postprint
Tipologia: Documento in Post-print
Licenza: Creative commons
Dimensione 9.5 MB
Formato Adobe PDF
9.5 MB Adobe PDF   Visualizza/Apri   Richiedi una copia
Baioletti2022_Chapter_ComparingBasinHoppingWithDiffe.pdf

non disponibili

Descrizione: Versione editoriale
Tipologia: Versione Editoriale (PDF)
Licenza: NON PUBBLICO - Accesso chiuso
Dimensione 909.36 kB
Formato Adobe PDF
909.36 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

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/20.500.12071/30505
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
social impact