| dc.description.abstract | The fine-tuning of the algorithms parameters, specially, in metaheuristics, is not always trivial and often is performed by ad hoc methods according to the problem under analysis. Usually, incorrect settings influence both in the algorithms performance, as in the quality of solutions. The tuning of metaheuristics requires the use of innovative methodologies, usually interesting to different research communities. In this context, this paper aims to contribute to the literature by presenting a methodology combining Statistical and Artificial Intelligence methods in the fine-tuning of metaheuristics. The key idea is a heuristic method, called Heuristic Oriented Racing Algorithm (HORA), which explores a search space of parameters, looking for candidate configurations near of a promising alternative, and consistently finds good settings for different metaheuristics. To confirm the validity of this approach, we present a case study for fine-tuning two distinct metaheuristics: Simulated Annealing (SA) and Genetic Algorithm (GA), in order to solve a classical task scheduling problem. The results of the proposed approach are compared with results yielded by the same metaheuristics tuned through different strategies, such as the brute-force and racing. Broadly, the proposed method proved to be effective in terms of the overall time of the tuning process. Our results from experimental studies reveal that metaheuristics tuned by means of HORA reach the same good results than when tuned by the other time-consuming fine-tuning approaches. Therefore, from the results presented in this study it is concluded that HORA is a promising and powerful tool for the fine-tuning of different metaheuristics, mainly when the overall time of tuning process is considered. | |
