Extremality of families of sets and set-valued optimization (Q6964776)

The paper continues the recent work of the same authors published in [\textit{N. D. Cuong} et al., Optimization 73, No. 12, 3593--3607 (2024; Zbl 1558.49027)], where a new extremality framework for collections of arbitrary families of sets was introduced. In the present work, the authors further investigate necessary optimality conditions for multiobjective and set-valued optimization problems in normed spaces. The analysis is based on the extremal principle, leading to the derivation of new extremal conditions. These results are applied to set-valued optimization problems of the form\N\[\N\mbox{minimize}\; F(x) \quad \mbox{subject to }\; x \in \Omega,\N\]\Nwhere \(F \colon X \rightrightarrows Y\) is a set-valued mapping between normed vector spaces, \(\Omega \subset X\), and the space \(Y\) is endowed with a general preference relation determined by an abstract level-set mapping \(L \colon Y \rightrightarrows Y\), as well as to more general problems with set-valued constraints. The results presented also cover the important special case in which \(Y\) is partially ordered by a nontrivial pointed convex cone, with optimality understood in the Pareto sense.