A unified approach for constrained extremum problems: image space analysis (Q382903): Difference between revisions

\noindent The paper deals with the following constrained extremum problem (CEP): \[ \min f(x)\text{ \;s.t. \;}x\in X,\text{ }G(x)\cap D\neq\emptyset, \] where \(f:\mathbb{R}^{n}\rightarrow\mathbb{R},\) \(G:\mathbb{R}^{n}\rightrightarrows \mathbb{R}^{m}\), \(X\) is a convex subset of \(\mathbb{R}^{n}\) and \(D\) is a closed, convex and pointed cone of \(\mathbb{R}^{m}\).\newline The image space analysis (ISA) is a unifying scheme for stuyding CEP problems. The ISA can be applied to any kind of problems that can be expressed under the form of the impossibility of a parametric system. The impossibility of such a system is reduced to the disjunction of two suitable subsets of the image space, which can be proved by showing that they lie in two disjoint level sets of a separating functional. When such a functional is linear, then it is said that they admit linear separation. A linear separation is said to be proper iff the two sets are not both contained in the separating hyperplane. A suitable subclass of separating functionals is said to be regular iff the separation obtained by means of a functional of such a subclass guarantees the disjunction between two suitable subsets of the image space.\newline The purpose of this paper is to apply the ISA to investigate a class of CEP, the constraining function of which is set-valued. The authors prove that (regular) separation in the image space is equivalent to the existence of saddle points of Lagrangian and generalized Lagrangian functions for CEP and the paper also gives Lagrangian type optimality conditions for CEP under suitable generalized convexity and compactness assumptions. The authors also present an exact penalty problem for CEP and show that it is equivalent to the existence of a regular linear separation under suitable generalized convexity and compactness assumptions.