Unboundedness of the images of set-valued mappings having closed graphs: Application to vector optimization

Vu Trung Hieu,Elisabeth Anna Sophia Köbis,Markus Arthur Köbis,Paul Hugo Schmölling
2024-09-22
Abstract:In this paper, we propose criteria for unboundedness of the images of set-valued mappings having closed graphs in Euclidean spaces. We focus on mappings whose domains are non-closed or whose values are connected. These criteria allow us to see structural properties of solutions in vector optimization, where solution sets can be considered as the images of solution mappings associated to specific scalarization methods. In particular, we prove that if the domain of a certain solution mapping is non-closed, then the weak Pareto solution set is unbounded. Furthermore, for a quasi-convex problem, we demonstrate two criteria to ensure that if the weak Pareto solution set is disconnected then each connected component is unbounded.
Optimization and Control
What problem does this paper attempt to address?