Dois enfoques para a resolução do problema de agrupamento

Abstract

In this paper, a study on two new approaches to the clustering problem is presented. Besides showing the detailed development of two proposed mathematical models, this research points to the importance of good modeling and the difference it can make in practice, contributing to the advancement in solving the problem and, consequently, to the field of mathematical programming in general. The distinguishing feature of the new models is that they are developed in a way that avoids the problem of non-differentiability and nonconvexity in their continuous relaxation. And the relevance of these new approaches are consolidated through the computational results developed as comparative experiments to show the strength of the proposed models in contrast to other known and studied models in the literature.