Star Death: a novel lightweight metaheuristic algorithm and its application for dynamic load-balancing in cluster computing

Optimization is a crucial principle in today's world, applied in various fields to increase profit and efficiency while reducing cost and time. However, solving optimization problems can be challenging, especially in dynamic environments where conditions are constantly changing. In the meantime...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Cluster computing Jg. 28; H. 9; S. 596
Hauptverfasser: Harifi, Sasan, Eghbali, Reza, Mirhosseini, Seyed Mohsen
Format: Journal Article
Sprache:Englisch
Veröffentlicht: New York Springer US 01.10.2025
Springer Nature B.V
Schlagworte:
Algorithms
Complexity
Computer Communication Networks
Computer Science
Dynamic loads
Effectiveness
Evolution
Genetic algorithms
Heuristic
Heuristic methods
Load balancing
Operating Systems
Optimization
Optimization techniques
Parameters
Problem solving
Processor Architectures
Lightweight metaheuristic
Metaheuristic
High-dimensional tests
Dynamic load-balancing
Star Death (SD) algorithm
Engineering problems
Physics-based algorithm
Optimization
ISSN:1386-7857, 1573-7543
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:Optimization is a crucial principle in today's world, applied in various fields to increase profit and efficiency while reducing cost and time. However, solving optimization problems can be challenging, especially in dynamic environments where conditions are constantly changing. In the meantime, metaheuristic methods are effective for solving large and complex optimization problems. Due to the presentation of several algorithms in the last two decades, each of which has high complexity and is difficult to understand, providing a lightweight algorithm has become a principle. This paper proposes a novel lightweight metaheuristic algorithm called the Star Death (SD) algorithm, which is inspired by the physical process of star death. The proposed algorithm aims to model the exact, regular, and optimal physical process of star death that can solve various problems. For this purpose, the SD algorithm employs an elite strategy that dynamically adjusts the range of exploration for better solutions. It also uses center-based sampling that emphasizes the center point's proximity to solutions, enhancing the optimizer's effectiveness. In this algorithm, the parameters are adjusted adaptively to enhance clarity and understanding of the parameter space. To prove application and robustness, the SD algorithm has been compared with 10 standard and popular metaheuristic algorithms. Based on this, 45 different benchmark test functions have been used. In addition, the algorithm has been tested and evaluated in high dimensions space. Also, it has been applied to 57 real-world CEC 2020 problems and six classic engineering problems. As a specific application, the SD algorithm is also used in solving the dynamic load-balancing problem. The results are generally indicative of the potential of the proposed algorithm to effectively solve complex optimization problems. The source codes of the SD algorithm are publicly available at https://github.com/harifi/SD .
Bibliographie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1386-7857
1573-7543
DOI:10.1007/s10586-025-05265-5