A multi-strategy improved electric eel foraging optimization algorithm: continuous and binary variants for solving optimization problems

Electric Eel Foraging Optimization (EEFO) algorithm is a metaheuristic inspired by the social predation behavior of electric eels. It incorporates interactions, resting, migration, and hunting activities to enhance search efficiency. Although EEFO is effective for optimization tasks, it is character...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:International journal of machine learning and cybernetics Ročník 16; číslo 9; s. 5985 - 6030
Hlavní autoři: Mostafa, Reham R., Khedr, Ahmed M., AL Aghbari, Zaher, Afyouni, Imad, Kamel, Ibrahim, Ahmed, Naveed
Médium: Journal Article
Jazyk:angličtina
Vydáno: Berlin/Heidelberg Springer Berlin Heidelberg 01.09.2025
Springer Nature B.V
Témata:
Algorithms
Artificial Intelligence
Classification
Complex Systems
Computational Intelligence
Control
Datasets
Design engineering
Energy consumption
Engineering
Evolutionary computation
Exploitation
Foraging behavior
Genetic algorithms
Global optimization
Heuristic
Heuristic methods
Mapping
Mechatronics
Mutation
Optimization algorithms
Optimization techniques
Original Article
Pattern Recognition
Physics
Rank tests
Robotics
Searching
Space exploration
Strategy
Systems Biology
Metaheuristic
Electric eel foraging optimization (EEFO)
Real-world optimization problems
Optimization
Differential evolution (DE)
ISSN:1868-8071, 1868-808X
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:Electric Eel Foraging Optimization (EEFO) algorithm is a metaheuristic inspired by the social predation behavior of electric eels. It incorporates interactions, resting, migration, and hunting activities to enhance search efficiency. Although EEFO is effective for optimization tasks, it is characterized by slower convergence rates and a tendency to fall into local optima in certain cases. To overcome these limitations, this paper proposes a Multi-strategy Improved Electric Eel Foraging Optimization (MIEEFO) that integrates three key strategies: adaptive tent chaotic mapping, Differential Evolution (DE) mutation strategy, and an enhanced solution technique based on the Fibonacci search technique (FSM). Firstly, MIEEFO employs an adaptive tent chaotic mapping strategy for initializing a uniformly distributed high-quality population for effective search space exploration. Secondly, a novel DE-based mutation strategy is introduced to balance the exploration and exploitation phases. Additionally, to enhance solution quality and mitigate the risk of local optima, an FSM-based improved solution technique is applied to refine the current optimal solution. To conduct a thorough assessment of MIEEFO’s global optimization capabilities, the established numerical challenge of the CEC’22 test suite is utilized. MIEEFO undergoes a comparative analysis with a range of modern, enhanced algorithms, employing the Wilcoxon signed-rank test and the Friedman test to integrate the results of these comparisons. The findings reveal that MIEEFO stands out for its superior optimization abilities, evidenced by its lowest average Friedman ranking of 1.37. MIEEFO consistently outperforms its rivals in most test scenarios, offering solutions that are both more precise and reliable. In addition, the application of MIEEFO is presented through five real-world constrained engineering design challenges, indicating its practical utility. These results highlight MIEEFO’s robust optimization capabilities and its potential for widespread application. Moreover, the proficiency of MIEEFO in managing discrete feature selection tasks is examined through tests on 17 datasets, in conjunction with ten established classification techniques and two advanced classification methods. The results confirm that MIEEFO achieved an average feature selection reduction of 72.59% across datasets while improving classification accuracy by up to 7.2% compared to competing methods.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1868-8071
1868-808X
DOI:10.1007/s13042-025-02609-w