Robust duality for robust efficient solutions in uncertain vector optimization problems (Q6104283)

The authors introduce the notion of second-order subdifferential for vector-valued and single-valued maps. Initially it is exhibited, through explicit examples and properties, the advantages of the new concept compared with previous ones, namely, second-order weakly subdifferential and second-order symmetric subdifferential, introduced in the articles of \textit{G. İnceoğlu} [Turk. J. Math. 45, No. 2, 955--960 (2021; Zbl 1507.49012)] and \textit{N. Q. Huy} and \textit{N. Van Tuyen} [J. Optim. Theory Appl. 171, No. 1, 27--44 (2016; Zbl 1353.49019)]. It is assured then the convexity of second-order subdifferentials and derived a global minimality condition for second-order subdifferentiable vector-valued maps. Under differentiability assumptions, homogeneity and supra-additivity properties of the second-order subdifferential are proven (Theorems 3.5, 3.6 and 3.7). Then the authors prove a necessary and sufficient optimality condition of robust efficient solutions to the given uncertain vector optimization problem (Theorem 5.1). The paper concludes firstly presenting, through second-order subdifferentials, a Wolfe type dual problem of the uncertain vector optimization problem, and secondly establishing robust weak/strong duality between the uncertain constrained vector optimization problem and its corresponding dual problem (Theorem 6.2).