A Comprehensive Review of Barnacles Mating Optimizer: Theoretical Foundation, Variants, Applications, and Future Research Directions

As real-world optimization problems become more complex, the development of sophisticated and robust algorithms has become essential. Consequently, researchers are focusing on advanced optimization methods that efficiently explore the feasible solution space. This involves designing new high-performance algorithms or enhancing existing meta-heuristic methods by integrating advanced evolutionary strategies. Barnacles Mating Optimizer (BMO) is an evolutionary-based meta-heuristic algorithm inspired by the mating behavior of barnacles, incorporating Hardy–Weinberg principles and the sperm-cast mechanism. Introduced in 2020, BMO has attracted significant attention and has been successfully applied across diverse fields due to its simple design, ease of implementation, high flexibility, and efficient convergence. Therefore, this review provides an overview and synthesis of studies employing BMO. It begins with an introduction to BMO, describing its natural inspiration and optimization framework, followed by a discussion of its core operational procedures and theoretical foundations. The paper then presents a comprehensive analysis of recent BMO variants, systematically categorizing them into modified, multi-objective, and hybrid versions. It also examines BMO’s diverse real-world applications, including power and control engineering, classification, image processing, wireless networks, forecasting, and signal processing. In addition, an updated performance evaluation of BMO is provided, comparing its effectiveness against recently published algorithms using the CEC2005 benchmark suite. Key strengths of BMO are highlighted, including its ability to balance exploration and exploitation, adaptability across problem domains, and its potential for hybridization with other optimization algorithms. Finally, potential enhancements and future research directions are outlined, including multi-objective variants, integration with deep learning, and parallel or distributed implementations.