Karush-Kuhn-Tucker optimality conditions for a class of robust optimization problems with an interval-valued objective function (Q2053413)

This article investigates an interval-valued robust optimization problem (IVROP), which is stated as follows \[ \left\{ \begin{array}{rcl} \inf f(x)&=&[f^{L}(x), f^{U}(x)]\\ \mathrm{ s. t.}&& g_i(x,v_i)\le 0,\,\forall\ v_i\in V_i, i=1,\ldots,m, \end{array} \right. \] where the objective function \(f\) is a locally Lipschitz interval-value function, and the constraint functions \(g_i\) are nonlinear and nonsmooth. The objective and constraint functions are defined on a Banach space. The sets \(V_i\) are assumed to be sequentially compact subsets of a topological space \(T_i\). This article establishes Karush-Kuhn-Tucker conditions in this nonlinear and nonsmooth setting. Indeed, Theorem 3.1 shows that, under certain additional assumptions, every non-dominated solution of the problem admits Fritz-John conditions in terms of generalized gradients, where the nonnegative multiplier is not trivial. Moreover, under an Extended Nonsmooth Mangasarian-Fromovitz constraint qualification, the optimality conditions become of KKT type (i.e., the multiplier of the objective function is positive).