Abstract

Computing intelligence is an important branch of artificial intelligence, and its techniques have been applied successfully in different research fields, such as mineral processing. These techniques include metaheuristic algorithms (MAs), which have been used to address difficult optimization problems. In this line, MAs have been proposed as an alternative to exact algorithms for designing flotation circuits, since the latter exhibit an extremely slow convergence even for moderately small instances. However, the MAs implemented in the literature to design flotation circuits are generally not validated or compared to other MAs to guarantee the quality of the solutions obtained. In this work, we present a methodology for developing, validating, and comparing MAs to design flotation circuits to full scale. The methodology is described considering the Tabu search algorithm, Differential Evolution Algorithm, and Particle Swarm Optimization, and their hybridizations. To improve the performance of MAs, we consider aspects such as competitive coevolution and one-dimensional and twodimensional chaotic maps. The results reveal that the methodology provides algorithms that exhibit a good relationship between the execution time and the quality of the results.