• Citer

Using experimental design to analyze stochastic local search algorithms for multiobjective problems

par Paquete, Luis;Stützle, Thomas ;López-Ibáñez, Manuel
Référence Operations research, computer science. Interface series, 39, page (325-344)
Publication Publié, 2007

Article révisé par les pairs

• Détails
• Contenu

Résumé :

Stochastic Local Search (SLS) algorithms can be seen as being composed of several algorithmic components, each playing some specific role with respect to overall performance. This article explores the application of experimental design techniques to analyze the effect of components of SLS algorithms for Multiobjective Combinatorial Optimization problems, in particular for the Biobjective Quadratic Assignment Problem. The analysis shows that there exists a strong dependence between the choices for these components and various instance features, such as the structure of the input data and the correlation between the objectives. © 2007 by Springer Science+Business Media, LLC.