Algumas contribuições para a otimização multiobjetivo via teoria dos cones

Abstract

This paper presents a characterization of solutions (efficient or pareto-optimal) for multiobjective optimization problems based on the calculation of sets tangent the theoretical discuss some elements of convex analysis, cones theory as well as multiobjective optimization elements needed for the formulation of the model. Furthermore, some classic multiobjectives methods are presented and classified into three types: priori, posteriori and interactive. From these elements, makes up the optimality conditions for general multiobjective problems, using sets of approximation of all of their issue viable set of the respective problem. At the end, the feasible set of the problem is defined by equality and inequality constraints and, with skill and regularity conditions, we obtain the Lagrange multipliers.