Algoritmos Quase-Newton para otimização multiobjetivo

Abstract

In this work, characterization are presented solutions for unconstrained multiobjective optimization for the cases of convex and non-convex function. The theoretical foundation of the convex case discusses a local solution obtained by solving a convex problem and some additional assumptions. For nonconvex case we show that the algorithm have a global convergence, in which the theoretical foundations ensure that curvature condition is obtained.