BHJO : A novel hybrid metaheuristic optimiser using Beluga Whale Optimisation, Honey Badger Algorithm and Artificial Jellyfish Search Optimiser for Optimisation Problems

Abstract

Metaheuristic optimization algorithms often have limitations in terms of their explo ration or exploitation capabilities. Therefore, relying on hybrid optimization algorithms has become essential to provide high-quality solutions. In this work, we propose a new hybrid optimization algorithm inspired by three recently developed algorithms, namely: Beluga Whale Optimization (BWO), Honey Badger Algorithm (HBA), and Artificial Jel lyfish Search Optimizer (JS). The HBA and BWO algorithms have demonstrated promis ing exploitation capabilities and stable exploration phases, while JS has exhibited global exploration capacity but lacks sufficient exploitation during the exploitation phase. By leveraging these characteristics, we have combined the three algorithms in the exploration phase and integrate HBA and BWO in the exploitation phase. Therefore, we introduce a novel approach to achieve a balance between exploration and exploitation. To en hance population diversity and effectively guide the search process, we employ Compound Opposition-Based Learning (COBL) technology.We thoroughly investigate and analyze the performance of the proposed BHJO algorithm by comparing it to the base algorithms HBA, BWO, JS, and four other modern algorithms. Our evaluation includes 20 standard benchmark problems, 20 hybrid test suites, and composite problems for unconstrained optimization from IEEE CEC2017, CEC2020, and CEC2021. The performance and be havior analysis are evaluated using the Friedman test, followed by the Dunn’s test to compare all the possible pairs of groups and identify the differences. The results demon strate that BHJO outperforms other algorithms in terms of balancing exploration and exploitation. The computational complexity is also evaluated.