An effective multi-objective artificial hummingbird algorithm with dynamic elimination-based crowding distance for solving engineering design problems

• Published in: Computer methods in applied mechanics and engineering Vol. 398; p. 115223
• Main Authors: Zhao, Weiguo, Zhang, Zhenxing, Mirjalili, Seyedali, Wang, Liying, Khodadadi, Nima, Mirjalili, Seyed Mohammad
• Format: Journal Article
• Language: English
• Published: Amsterdam Elsevier B.V 01.08.2022; Elsevier BV
• Subjects: Algorithms; Archives & records; Artificial hummingbird algorithm; Convergence; Convergence and diversity; Crowding; Design engineering; Design optimization; Dynamic elimination-based crowding distance; Engineering design problems; Heuristic methods; Multi-objective optimization; Multiple objective analysis; Non-dominated sorting; Pareto optimization; Pareto optimum; Source code; Dynamic elimination-based crowding distance; Multi-objective optimization; Artificial hummingbird algorithm; Engineering design problems; Convergence and diversity; Non-dominated sorting
• ISSN: 0045-7825; 1879-2138
• DOI: 10.1016/j.cma.2022.115223

Artificial hummingbird algorithm (AHA) is a recently developed bio-based metaheuristic and it shows superior performance in handling single-objective optimization problems. Despite the merit, this algorithm can only solve problems with one objective. To solve complex multi-objective optimization problems, including engineering design problems, a multi-objective AHA (MOAHA) is developed in this study. In MOAHA, an external archive is employed to save Pareto optimal solutions, and a dynamic elimination-based crowding distance (DECD) method is developed to maintain this archive to effectively preserve the population diversity. In addition, a non-dominated sorting strategy is merged with MOAHA to construct a solution update mechanism, which effectively refines Pareto optimal solutions for improving the convergence of the algorithm. The superior results over 7 competitors on 28 benchmark functions in terms of convergence, diversity and solution distribution are demonstrated with a suite of comprehensive tests. The MOAHA algorithm is also applied to 5 real-world engineering design problems with multiple objectives, demonstrating its superiority in handling challenging real-world multi-objective problems with unknown true Pareto optimal solutions and fronts. The source code of MOAHA is publicly available at https://ww2.mathworks.cn/matlabcentral/fileexchange/113535-moaha-multi-objective-artificial-hummingbird-algorithm and https://seyedalimirjalili.com/aha. •Multi-objective artificial hummingbird algorithm (MOAHA) is proposed.•Dynamic elimination-based crowding distance is used to produce a well-distributed PF.•A solution update mechanism using NDS is employed to update non-dominated solutions.•A visit table maintenance mechanism is employed in three foraging strategies.•It shows superior results over 7 competitors on 28 functions and 5 engineering cases.