Hyperheuristics for a dynamic-mapped multi-objective Island-based model

Abstract

This work presents a set of improvements and a performance analysis for a previously designed multi-objective optimisation parallel model. The model is a hybrid algorithm that combines a parallel island-based scheme with a hyperheuristic approach in order to grant more computational resources to those schemes that show a more promising behaviour. The main aim is to raise the level of generality at which most current evolutionary algorithms operate. This way, a wider range of problems can be tackled since the strengths of one algorithm can compensate for the weaknesses of another. A contribution-based hyperheuristic previously presented in the literature is compared with a novel hypervolume-based hyperheuristic. The computational results obtained for some tests available in the literature demonstrate the superiority of the hypervolume-based hyperheuristic when compared to the contribution-based hyperheuristic and to other standard parallel models. © 2009 Springer Berlin Heidelberg.