Analysing the adaptation level of parallel hyperheuristics applied to mono-objective optimisation problems

Abstract

Evolutionary Algorithms (eas) are one of the most popular strategies for solving optimisation problems. One of the main drawbacks of eas is the complexity of their parameter setting. This setting is mandatory to obtain high quality solutions. In order to deal with the parameterisation of an ea, hyperheuristics can be applied. They manage the choice of which parameters should be applied at each stage of the optimisation process. In this work, an analysis of the robustness of a parallel strategy that hybridises hyperheuristics, and parallel island-based models has been performed. Specifically, the model has been applied to a large set of mono-objective scalable benchmark problems with different landscape features. In addition, a study of the adaptation level of the proposal has been carried out. Computational results have shown the suitability of the model with every tested benchmark problem. © 2011 Springer-Verlag Berlin Heidelberg.