Automatic Tuning of GRASP with Path-Relinking in data clustering with F-R ace and iterated F-Race

Authors

Júlio Cesar da Silva, Universidade Federal de ItajubáFollow
Rafael de Magalhães Dias Frinhani, Universidade Federal de ItajubáFollow
Ricardo Martins de Abreu Silva, Universidade Federal de PernambucoFollow
Geraldo Robson Mateus, Universidade Federal de Minas GeraisFollow

Document Type

Article

Publication Date

5-2015

Keywords

Automatic Tunig, I/F-Race, GRASP, Path-Relinking, Data Clustering

Abstract

In studies that use metaheuristics although the input parameters directly influence the performance of the algorithm its definition is mostly done manually raising questions about the quality of the results. This paper aims to apply the F/I-Race in the self parameterization of GRASP with Path-Relinking in the data clustering in order to obtain better results than the manually tuned algorithms. Experiments performed with five data sets showed that the use of I/F-race contributed to achievement best results than manual tuning.

Comments

This paper is in Portuguese (Auto-parametrização do GRASP com Path-Relinking no agrupamento de dados com F-Race e iterated F-Race)

Recommended Citation

da Silva, Júlio Cesar; Frinhani, Rafael de Magalhães Dias; Silva, Ricardo Martins de Abreu; and Mateus, Geraldo Robson, "Automatic Tuning of GRASP with Path-Relinking in data clustering with F-R ace and iterated F-Race" (2015). Proceedings of the XI Brazilian Symposium on Information Systems (SBSI 2015). 94.
https://aisel.aisnet.org/sbis2015/94