Research Area:  Metaheuristic Computing

Abstract:

This study develops a Multi-Objective Jellyfish Search (MOJS) algorithm to solve engineering problems optimally with multiple objectives. Lévy flight, elite population, fixed-size archive, chaotic map, and the opposition-based jumping method are integrated into the MOJS to obtain the Pareto optimal solutions. These techniques are employed to define the motions of jellyfish in an ocean current or a swarm in multi-objective search spaces. The proposed algorithm is tested on 20 multi-objective mathematical benchmark problems, and compared with six well-known metaheuristic optimization algorithms (MOALO, MODA, MOEA/D, MOGWO, MOPSO, and NSGA-II). The results thus obtained indicate that the MOJS finds highly accurate approximations to Pareto-optimal fronts with a uniform distribution of solutions for the test functions. Three constrained structural problems (25-bar tower design, 160-bar tower design, and 942-bar tower design) of minimizing structural weight and maximum nodal deflection were solved using MOJS. The visual analytics demonstrates the merits of MOJS in solving real engineering problems with best Pareto-optimal fronts. Accordingly, the MOJS is an effective and efficient algorithm for solving multi-objective optimization problems.

Keywords:  
multi-Objective
jellyfish Search
engineering problem
chaotic map
engineering problem
best Pareto-optimal front

Author(s) Name:  Jui-Sheng Chou, Dinh-Nhat Truong

Publisher name:  Elsevier

DOI:  10.1016/j.chaos.2020.109738

Volume Information:  Volume 135

Paper Link:   https://www.sciencedirect.com/science/article/abs/pii/S0960077920301405