5 - MOGROM: Multiobjective Golden Ratio Optimization Algorithm

Abstract

This study introduces the multiobjective version of a very recently proposed metaheuristic optimization algorithm known as the Golden Ratio Optimization Method (GROM). The proposed method, named Multiobjective GROM (MOGROM), uses an external Repository matrix for storing the best obtained Pareto Front that is achieved by nondominated sorting of the whole population. The members of the Repository are also sorted based on a crowding distance concept to calculate the best and the worst individuals corresponding to the highest and lowest crowding distance, respectively. To evaluate the performance of the proposed algorithm, 22 multiobjective standard test functions with different Pareto Front shapes and search spaces are employed. The obtained results using the MOGROM are compared to those of five well-known multiobjective optimization algorithms including NSGA-II, MOPSO, MOALO, MOGWO, and MOLAPO. Four different criteria, namely, Generational Distance (GD), Inverted GD, Metric of Spread, and Metric of Spacing, are also employed for the sake of results’ comparison. The comparisons demonstrate the excellent performance of the proposed method in solving different types of multiobjective problems.