Shake them all! Rethinking Selection and Replacement in MOEA/D
Gauvain Marquet, Bilel Derbel, Arnaud Liefooghe, El-Ghazali Talbi · 2014
Abstract. In this paper, we build upon the previous efforts to en-hance the search ability of Moead (a decomposition-based algorithm), by investigating the idea of evolving the whole population simultane-ously at once. We thereby propose new alternative selection and re-placement strategies that can be combined in different ways within a generic and problem-independent framework. To assess the performance of our strategies, we conduct a comprehensive experimental study on bi-objective combinatorial optimization problems. More precisely, we con-sider ρMNK-landscapes and knapsack problems as a benchmark, and experiment a wide range of parameter configurations for Moead and its variants. Our analysis reveals the effectiveness of our strategies and their robustness to parameter settings. In particular, substantial im-provements are obtained compared to the conventional Moead. 1