MOAHHGSO-w: A new hybrid algorithm for solving multi-objective optimization problems

Abstract

To effectively address multi-objective optimization problems, this study proposes a novel hybrid optimization approach, referred to as MOAHHGSO-w. This method integrates the characteristics of the multi-objective Weighted Sum (WS) algorithm with two recently developed single-objective optimization techniques: the Artificial Hummingbird Algorithm (AHA) and Henry Gas Solubility Optimization (HGSO). Furthermore, MOAHHGSO-w utilizes three distinct population archives to store the non-dominated solutions identified during the search process. To enhance solution diversity, the algorithm incorporates food source selection and leader selection mechanisms, both of which rely on the crowding distance metric. To achieve an optimal balance between solution diversity and convergence toward the Pareto-optimal set, the proposed method employs a combination of crowding distance and epsilon dominance. The algorithm is evaluated on four real-world multi-objective engineering design challenges. Its performance is benchmarked against several well-established multi-objective optimization methods. To validate the experimental findings, the study employs two widely recognized performance metrics: Inverted Generational Distance (IGD) for assessing convergence and Spacing (Sp) for evaluating solution diversity. The statistical analysis confirms that the proposed MOAHHGSO-w algorithm outperforms the compared optimization techniques in terms of efficiency and effectiveness.