Authors:
Saman M. Almufti, Amira Bibo Sallow
Addresses:
Department of Information Technology, Akre Technical College of Informatics, Akre University for Applied Sciences, Akre, Duhok, Iraq. Department of Computer Science, College of Science, Knowledge University, Erbil, Kurdistan Region, Iraq. Department of Information Technology, Duhok Technical College, Duhok Polytechnic University, Duhok, Kurdistan Region, Iraq.
Abstract:
Metaheuristic algorithms have become indispensable for solving complex optimisation problems characterised by multimodality, non-separability, and high dimensionality. This study presents a rigorous comparative analysis of three notable metaheuristics: the classical Particle Swarm Optimisation (PSO), the Grey Wolf Optimiser (GWO), and the Enhanced Social Network Search (ESNS) algorithm. Using a suite of ten standardised benchmark functions (F1–F10), encompassing both unimodal and multimodal landscapes, we evaluate each algorithm's convergence behaviour, robustness, and scalability. The functions range from the simple Sphere function (F1) to the highly complex Shifted Rotated Griewank function (F10), offering a comprehensive assessment framework. Experimental results, expressed in terms of mean and standard deviation over multiple runs, demonstrate the superiority of ESNS in handling high-dimensional and transformed landscapes, followed by GWO. PSO, while efficient in simpler domains, exhibits reduced performance in complex scenarios due to premature convergence. These findings underscore the importance of adaptive mechanisms and statistically guided exploration in modern optimisation. The study offers valuable insights into algorithm selection for real-world engineering and scientific applications.
Keywords: Metaheuristic Optimization; Particle Swarm Optimization (PSO); Grey Wolf Optimiser (GWO); Enhanced Social Network. Search (ESNS); Benchmark Functions; Swarm Intelligence; Convergence Analysis.
Received: 26/08/2024, Revised: 13/11/2024, Accepted: 30/12/2024, Published: 07/06/2025
DOI: 10.64091/ATICS.2025.0000131
AVE Trends in Intelligent Computing Systems, 2025 Vol. 2 No. 2 , Pages: 62-76