Proceedings of The 5th International Conference on Innovation in Science and Technology
Performance Assessment of Multi-Objective Optimization Algorithms on Large-Scale Problems
T. Sağ , A. Özkış
Multiobjective optimization (MO) has been an attractive field in recent decades and many different algorithms have been proposed to solve MO problems. Although a lot of studies were focused to small-scale problems, real-world optimization problems frequently consist a large number of decision variables. In this study, the capabilities of the MO algorithms on large-scale optimization problems are investigated. For this purpose, four different techniques (NSGAII, MOCell, IBEA, and MOEA/D) known as the-state-of-art algorithms in the specialized literature are applied to solve a set of standard benchmark problems called ZDT functions for 10, 50, 100, 200, 500, and 1000 variable instances. Also, the values with standard deviations of four performance indicators (HV, SP, ε, and IGD) are calculated to promote the research. The experimental results have demonstrated that MOCell is generally able to reach the superior results than the other algorithms under the all conditions.