Comparing and Ranking of Meta-Heuristic Algorithms Using Group Decision Making Methods

In recent years, meta-heuristic algorithms and their application in solving complicated, nonlinear and NP-hard problems have dramatically increased, while new algorithms have constantly being introduced. In this research, with the aim of ranking meta-heuristic algorithms, using group decision making techniques (different from other research in this field), 5 algorithms including: GA, PSO, ABC, SFLA and ICA by 15 standard test functions, and considering 2 attribute: "mean of answers" and "run time", have been compared. Then they are ranked by 3 group decision making methods including: "Cook and Seiford", "Condorcet" and "Dodgson". In addition, as in ranking by "Condorcet" and "Dodgson" methods, sometimes some options posit the same rank, therefore, in this study; we presented a proposal to overcome the limitation. Then the algorithms with these proposed methods were ranked. Finally, the overall ranking is done using an allocation model our results show that the overall ranking is as follows, respectively: PSO, ICA, GA, ABC and SFLA.