Research report: Hooded seal optimization algorithm: a new metaheuristic to optimize engineering problem

A new single-objective swarm-based evolutionary metaheuristic algorithm inspired by the behavior of hooded seals has been introduced to address the persistent computational challenge where optimization algorithms struggle to escape local optima in high-dimensional problems. Developed by researchers Solav J. Omar and Chnoor M. Rahman and published in Neural Computing and Applications, the proposed computational method is designated as Hooded Seal optimization, or HoSeO. The core objective of the algorithm is to overcome the problem of local optima entrapment by establishing a favorable balance between exploration and exploitation across the search domain.

The underlying mathematical architecture of the Hooded Seal optimization algorithm relies on three principal computational procedures, each governed by dynamic parameter control to regulate the search behavior over time. The first procedure is a specialized mating phase that functions through proximity-strength pairing between individuals.

To prevent loss of diversity while driving the search toward optimal solutions, HoSeO incorporates a unique population update strategy designed to preserve both solution quality and variety within the swarm.

The computational efficacy of HoSeO was subjected to rigorous empirical evaluation across multiple widely recognized benchmark suites, encompassing a total of 45 mathematical test functions alongside six real-world problems. Specifically, the experimental framework tested the algorithm on 23 classical benchmark functions, 10 CEC-2019 test suite functions, and 12 CEC-2022 benchmark functions.

To establish its comparative standing within the field of evolutionary computation, the performance of HoSeO was evaluated against several well-known and competitive metaheuristic algorithms. The baseline methods selected for comparison included Genetic Algorithm (GA), Multi-Verse Optimizer (MVO), White Shark Optimizer (WSO), Learner Performance-based Behavior (LPB), Dragonfly Algorithm (DA), Linear Population Size Reduction Success-History Based Parameter Adaptation Differential Evolution (L-SHADE), and Raindrop Algorithm (RD).

Experimental results demonstrated that HoSeO consistently delivered superior optimization performance across the evaluated test sets, outperforming or matching the competitive algorithms on 17 out of the 23 classical benchmark functions. In modern benchmark challenges, HoSeO achieved dominance in 7 out of the 10 functions comprising the CEC-2019 test suite and ranked among the top-performing methods in 9 out of the 12 CEC-2022 benchmark functions. Furthermore, evaluation across six real-world problems highlighted the algorithm’s strong capability to resolve complex and diverse practical optimization tasks, with the statistical significance and validity of the overall findings confirmed through rigorous statistical validations.

Subject of Research: Swarm-based evolutionary metaheuristic optimization

Article Title: Hooded seal optimization algorithm: a new metaheuristic to optimize engineering problem

Article References: J.Omar, S., & Rahman, C. M. (2026). Hooded seal optimization algorithm: a new metaheuristic to optimize engineering problem. Neural Computing and Applications, 38(21), Article 875. https://doi.org/10.1007/s00521-026-12603-0

Image Credits: AI Generated

DOI: 10.1007/s00521-026-12603-0

Keywords: optimization algorithms, metaheuristics, swarm intelligence, Hooded Seal optimization, evolutionary computation, engineering optimization, Lévy flight, benchmark functions, Hooded, seal, optimization, algorithm