Higher-order optimality conditions for strict and weak efficient solutions in set-valued optimization (Q291997)

Extending the higher-order derivative of \textit{M. Studniarski} [SIAM J. Control Optim. 24, 1044--1049 (1986; Zbl 0604.49017)] for set-valued maps to the (cone-)epigraph of the set-valued map \(F\), \(m\)th-order necessary (Th. 4.1, weak) and sufficient (Th. 4.2, strict) optimality conditions are proved for the following problem. Determine a (weak, strict) efficient point of \(F(x)\) w.r.t. a pointed closed convex cone \(C\subset Y\) with nonempty interior where \(x\) belongs to some set \(S\neq \emptyset\) of the normed space \(X\) such that \(G(x)\cap (-D)\neq\emptyset\). \(F:X\rightarrow 2^Y\) and \(G:X\rightarrow 2^Z\) are set-valued mappings, \(X,Y,Z\) are normed spaces and \(D\) is a pointed closed convex cone in \(D\) with nonempty interior. For all derived conditions the existence of the used higher-order set derivatives is assumed. The necessary multiplier condition is derived under invexity of \(S\), convex likeness of \((F-y_0,G)\) in some local sense w.r.t. the solution \((x_0,y_0)\). The sufficient multiplier condition of \(m\)th order for strict efficiency needs the following stability condition of order \(m\) at \((x_0,y_0)\) for \(F\): There is some \(\lambda >0\) and some neighborhood \(U\) of \(x\) such that for all \(x\in U\setminus\{x_0\}\) holds \(F(x)\subseteq \{y_0\}+\lambda\|x-x_0\|^m B_Y\) where \(B_Y\) is the closed unit ball of \(Y\).