A new selection strategy for multi objective genetic algorithm: MultiMoora Rank selection

Abstract

Genetic algorithms are metaheuristic methods that provide satisfactory solutions for complex problems within acceptable time periods. In genetic algorithms, the selection strategy considerably affects the quality of the solution. In this study, MultiMoora Rank Selection (MMRS) based on MultiMoora method, which is one of the multi criteria decision making methods, has been developed as a new selection strategy to increase the performance of Multi Objective Genetic Algorithm (MOGA). The performance of the method was tested using flow shop scheduling problems. The results obtained with the MultiMoora Rank Selection were evaluated by comparing with the results obtained using the same problem with the selection methods that are widely used for genetic algorithms, such as the Roulette Wheel Selection, Linear Rank Selection and Tournament Selection methods. The results showed that the proposed MultiMoora Rank Selection method outperformed the compared methods.