Limite Superior para o Problema da Diversidade Máxima

Abstract

We apply the t-linearization to the straightforward 0 1 quadratic model for the maximum diversity problem. We computationally compare the obtained model with two other linear formulations from the literature. Computational experiments show that the t-linearization generates, on average, fewer constraints than the formulations, in addition to achieving a tighter upper bound in all used instances. On the other hand, one of the formulations can be slightly faster in obtaining the bounds.