A new Fenchel dual problem in vector optimization (Q841208)

Duality theory gives dual characterizations for the optimal solutions of a primal problem. Its extensions for multiobjective optimization problems go in different directions. One of them is the extension of the classical Lagrange duality approach. In another one a vector dual using the Fenchel-Lagrange dual of a scalar convex optimization problem is constructed. The second approach is developed further. The primal problem treated in this paper has as objective function the sum of a vector function with another one, which is the composition of a vector function with a linear operator. The authors attach to the primal vector problem a scalarized optimization problem and consider its scalar Fenchel dual problem. Inspired by the formulation of the scalarized dual they define a new vector dual problem. Weak and strong duality is proved and compared with two other duals -- one inspired by Breckner and Kolumban, the second constructed by Jahn. The image sets of these three duals are strictly connected. Some properties and inclusions are proved. It is also shown that their sets of maximal elements are equal.