Impact of Boundary Control Methods on Bound-constrained Optimization Benchmarking

Result code in IS VaVaI

<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F70883521%3A28140%2F23%3A63563658" target="_blank" >RIV/70883521:28140/23:63563658 - isvavai.cz</a>

Result on the web

<a href="https://dl.acm.org/doi/10.1145/3583133.3595849" target="_blank" >https://dl.acm.org/doi/10.1145/3583133.3595849</a>

DOI - Digital Object Identifier

<a href="http://dx.doi.org/10.1145/3583133.3595849" target="_blank" >10.1145/3583133.3595849</a>

Result language

angličtina

Original language name

Impact of Boundary Control Methods on Bound-constrained Optimization Benchmarking

Original language description

Despite initial indifference towards boundary control methods (BCM) in the context of metaheuristic algorithm design, benchmarking, and execution, our research demonstrates their critical importance. This study investigates how the choice of a particular BCM can profoundly influence the performance of competitive algorithms. We analyzed the top three algorithms from the 2017 and 2020 IEEE CEC competitions, posing the following question: Could a change in BCM usage alter an algorithm&apos;s overall performance and, consequently, its ranking among competitors? Our findings reveal that paying attention to BCMs can lead to significant improvements. The experiments revealed that BCM selection can significantly impact an algorithm&apos;s performance and, in some instances, its competition rank. However, most authors omitted to mention the implemented BCM, resulting in poor reproducibility and deviating from recommended benchmarking practices for metaheuristic algorithms. The conclusion is that the BCM should be considered another vital metaheuristics input variable for unambiguous reproducibility of results in benchmarking and for a better understanding of population dynamics.

Czech name

—

Czech description

—

Type

D - Article in proceedings

CEP classification

—

OECD FORD branch

10201 - Computer sciences, information science, bioinformathics (hardware development to be 2.2, social aspect to be 5.8)

Project

—

Continuities

S - Specificky vyzkum na vysokych skolach

Publication year

2023

Confidentiality

S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Article name in the collection

GECCO 2023 Companion - Proceedings of the 2023 Genetic and Evolutionary Computation Conference Companion

ISBN

979-840070120-7

ISSN

—

e-ISSN

—

Number of pages

2

Pages from-to

25-26

Publisher name

Association for Computing Machinery, Inc

Place of publication

New York

Event location

Lisbon

Event date

Jul 15, 2023

Type of event by nationality

WRD - Celosvětová akce

UT code for WoS article

—