Um algoritmo matemático para programação vetorial

Abstract

This paper presents an algorithm that uses the descent method to for solve a vector optimization problem unconstrained multiobjective where the functions considered are continuously differentiable. It will also be a study on the theoretical foundations, namely: elements of convex analysis, induced partial order by a generic cone K, as well as multi-objective and vectorial programming fundamentals, required for formulation of the mathematical model. To calculate the direction of descent, an auxiliary function strongly convex and is used for the step size, the Armijo rule type. It is shown that the whole point of accumulation of the generated sequence the algorithm is K-critical for the vector.