Nonlinear scalarizations in set optimization with variable ordering structures and applications (Q6944564)

The set optimization problem (P) \(F(x)\rightarrow \min\), \(x\in M\) is considered where \(X,Y\) are normed spaces and \(F:X\rightrightarrows Y\) is a set valued mapping with non empty values. The set valued mapping \(K:Y\rightrightarrows Y\) has as values pointed convex closed cones with non empty interior and defines in \(Y\) a lower type variable order structure \(A\preccurlyeq^K_l B:\Leftrightarrow B \subset \bigcup_{a\in A}(a+K(a))=: A^K\), a so called \(l\)-type set less order relation. Replacing \(K(a)\) by \(\operatorname{int} K(a)\) the strict less order relation \(\prec^K_l\) is given. \(x_0\in M\) is (weakly) optimal w.r.t. (P) if \(F(x)(\prec^K_l) \preccurlyeq^K_l F(x_0) \Longrightarrow F(x_0)(\prec^K_l) \preccurlyeq^K_l F(x)\). For some element \(e \in \bigcap_{y\in Y}\operatorname{int} K(y)\) the scalarization function \(\xi_x(A):=\inf\left\lbrace t\in \mathbb{R}: A \preccurlyeq^K_l te +F(x)\right\rbrace\) is suggested. Its monotony and order representing properties are shown and used for the characterization of optimal and weakly optimal solutions. A lot of very illustrative examples underline the importance of the used assumptions. Some results in the literature are generalized.