On smoothness properties of optimal value functions at the boundary of their domain under complete convexity (Q2247981): Difference between revisions

The authors consider the parametric optimization problem \(P(t)\): min\(_{x} f(x,t) \) subject to \(x\in M(t)\), where \(f: \mathbb R^r\times \mathbb R^n\to \mathbb R\) is a convex function and \(M: \mathbb R^r\rightrightarrows \mathbb R^n\) is a set-valued mapping with closed and convex graph. They give some continuity properties of the optimal value function and sufficient conditions for inner semi-continuity of the mapping \(M\). The main contribution is concerned with directional differentiability of the optimal value function at boundary points of its domain when \(M\) is given by a system of convex inequalities. They provide formulas to practically compute directional derivatives at these points by using Karush-Kuhn-Tucker multipliers and the Lagrange function.