Sardine Optimization Algorithm with Agile Locality and Globality Strategies for Real Optimization Problems

How to steadily find satisfactory solutions for high-dimensional multimodal and composition optimization problems is still a challenging issue. To fight against this pain-point problem, we propose sardine optimization algorithm (SOA) with agile locality and globality strategies for real optimization...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Arabian journal for science and engineering (2011) Ročník 48; číslo 8; s. 9787 - 9825
Hlavní autoři: Zhang, HongGuang, Tang, MengZhen, Liu, YuanAn, Li, Xiang
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.08.2023
Springer Nature B.V
Témata:
Algorithms
Composition
Engineering
Heuristic methods
Humanities and Social Sciences
multidisciplinary
Operators
Optimization
Optimization algorithms
Research Article-Computer Engineering and Computer Science
Sardines
Science
Searching
Swarm intelligence
Meta-heuristic algorithm
Evolutionary algorithms
Agile strategies
Optimization
ISSN:2193-567X, 1319-8025, 2191-4281
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:How to steadily find satisfactory solutions for high-dimensional multimodal and composition optimization problems is still a challenging issue. To fight against this pain-point problem, we propose sardine optimization algorithm (SOA) with agile locality and globality strategies for real optimization problems. Inspired by the survival philosophy of sardines, SOA simulates the transformation, migration, reproduction, elimination, and spread of sardines. As a varied-population-size optimization algorithm, the features of SOA are summarized as two key points. (i) Agile locality and globality strategies use adjacent and corresponding period ratios to control the local and global search behaviors. To our best knowledge, these strategies are a new technical road to balance exploration and exploitation efforts. (ii) Besides, SOA uses unique search operators based on the center movement of sardine schools. Specifically, when the center of one sardine school moves in these search operators, all sardines in this school also move in the same direction. This mobility style is different from most meta-heuristic algorithms, as far as we know. We used all unconstrained optimization problems in the CEC2013 test suite and six real-world constrained optimization problems as our benchmarks. SOA outperforms eight algorithms (like the two winning algorithms of CEC2013 and CEC2014), especially for high-dimension multimodal and composition optimization problems. For instance, Wilcoxon results between SOA and the two winning algorithms of CEC2013 and CEC2014 for 90-dimensional unconstrained optimization problems of the CEC2013 test suite are 16:10 and 18:8. The coding of SOA can be downloaded from “ https://github.com/1654402787/SOA ” (unzip password: soasoasoa1357).
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:2193-567X
1319-8025
2191-4281
DOI:10.1007/s13369-022-07350-y