Automatic Configuration of Metaheuristics for Solving the Quadratic Three-Dimensional Assignment Problem Using Irace.

Metaheuristic algorithms are traditionally designed following a manual and iterative algorithm development process. The performance of these algorithms is, however, strongly dependent on their correct tuning, including their configuration and parametrization. This is labour-intensive, error-prone, difficult to reproduce and explores only a limited number of design alternatives. To overcome manual tuning, the automatic configuration of algorithms is a technique that has shown its efficiency in finding performance-optimizing settings of parameters. This paper contributes to overcoming the challenge of automatically configured metaheuristics using the iterated racing for automatic algorithm configuration irace applied to the quadratic three-dimensional assignment problem. In particular, we use particle swarm optimization (PSO), a tabu search (TS), an iterated local search (ILS) and two hybrid algorithms PSO-TS and PSO-ILS. Of these algorithms, the tabu search algorithm and the PSO-ILS worked the best. The results show that the algorithm automatic configuration enables identifying an ideal tuning of the parameters and reaching better results when compared to a manual configuration, in similar execution time.