par Perez Caceres, Leslie 
;López-Ibáñez, Manuel ;Stützle, Thomas
Référence Lecture notes in computer science, 8600, page (37-48)
Publication Publié, 2014
;López-Ibáñez, Manuel ;Stützle, Thomas Référence Lecture notes in computer science, 8600, page (37-48)
Publication Publié, 2014
Article révisé par les pairs
| Résumé : | The irace package implements a flexible tool for the automatic configuration of algorithms. However, irace itself has specific parameters to customize the search process according to the tuning scenario. In this paper, we analyze five parameters of irace: the number of iterations, the number of instances seen before the first elimination test, the maximum number of elite configurations, the statistical test and the confidence level of the statistical test. These parameters define some key aspects of the way irace identifies good configurations. Originally, their values have been set based on rules of thumb and an intuitive understanding of the configuration process. This work aims at giving insights about the sensitivity of irace to these parameters in order to guide their setting and further improvement of irace. |
