Multi-Objective Passing Vehicle Search algorithm for structure optimization

• Published in: Expert systems with applications 2021-05, Vol.169, p.114511, Article 114511
• Main Authors: Kumar, Sumit, Tejani, Ghanshyam G., Pholdee, Nantiwat, Bureerat, Sujin
• Format: Article
• Language: English
• Subjects: Algorithms; Constraint optimization; Design optimization; Discrete variables; Metaheuristic; Multi-objective optimization; Multiple objective analysis; Performance evaluation; Rank tests; Search algorithms; Structural design
• ISSN: 0957-4174; 1873-6793
• DOI: 10.1016/j.eswa.2020.114511

•A novel metaheuristic called Multi-objective Passing Vehicle Search algorithm (MOPSV)•MOPSV performance validated with 5 multi-objective truss optimization problems.•Performance comparison with other state-of-the-art metaheuristics.•MOPSV has a good balance of diversification and intensification of search.•Performance assessment reveals that MOPVS is a very powerful algorithm. A novel Multi-Objective Passing Vehicle Search (MOPVS) algorithm is proposed for structural design optimization. MOPVS is inspired by the two-lane highway passing vehicle mechanism. This multi-objective version is modified and further improved from the single-objective version of passing vehicle search through a Pareto dominance-based approach. For performance evaluation of MOPVS, five daunting benchmark structural design problems have been used. Two conflicting objectives i.e. structure weight minimization and minimization of maximum nodal displacement along with discrete design variables have been considered to ensure its real-world applications. For fitness and efficiency evaluation of the proposed algorithm, the results obtained from the new algorithm are compared with four other state-of-the-art multi-objective algorithms. Moreover, two performance indicators test called Hypervolume and Spacing-to-Extent were performed for the rigorous evaluation of the performance and feasibility of the proposed algorithm. The findings demonstrate the superiority of the MOPVS algorithm over the others while the potential to find a non-dominated solution set with diverse individual solutions. Present work considers the Friedman’s rank test for the statistical investigation of the experiment work. The solutions and convergence behavior achieved by MOPVS show its high efficiency in solving challenging design problems.