%0 Conference Proceedings
%4 sid.inpe.br/mtc-m21c/2019/08.01.16.11
%2 sid.inpe.br/mtc-m21c/2019/08.01.16.11.57
%T A heuristic oriented racing algorithm for the fine-tuning of metaheuristics
%D 2019
%A Barbosa, Eduardo Batista de Moraes,
%A Senne, Edson L. F.,
%@affiliation Instituto Nacional de Pesquisas Espaciais (INPE)
%@affiliation Universidade Estadual Paulista (UNESP)
%@electronicmailaddress eduardo.barbosa@inpe.br
%@electronicmailaddress edson.senne@unesp.br
%B Metaheuristics International Conference
%C Cartagena, Colombia
%8 28-31 July
%S Proceedings
%X The metaheuristics have become a powerful tool to solve real-world optimization problems. Its ease adaptability, usually demands effort to correctly define its components (e.g.: problem representation, neighborhood structure, etc.) and parameters to achieve their best performance. Thus, this paper aims to present an approach on the fine-tuning of metaheuristics combining Design of Experiments and Racing algorithms. The key idea is a heuristic method, which explores a search space of parameters looking for candidate configurations near of a promising alternative and consistently finds the good ones. To confirm this approach, we present a case study for fine-tuning a VNS metaheuristic on the classical Traveling Salesman Problem, and compare its results against a well established racing method. In general, our approach proved to be effective in terms of the overall time of the tuning process.
%@language en
%3 Barbosa_heuristic.pdf