On proto-differentiability of generalized perturbation maps (Q852846): Difference between revisions

In this very important paper, the authors deal with the sensitivity analysis for the optimization problems and the variational inequalities. Here, the parameterized convex programming problems and the parameterized variational inequality problems are presented by using the generalized equation with the following form: \(z\in F(x,y)+N_D(y)\), where \(F\) is a multifunction from \(R^d\times R^n\) to \(R^n\), and \(N_D(y)\) denotes the normal cone to \(D\) (\(D\subset R^n\) is a convex set). To obtain general results for multifunctions, the generalized perturbation map \(G:R^d \times R^n = R^n\), given by \(G(x,z)=\{y\in D:z-F(x,y) \in N_D (x,y)\}\) is studied. Main result: Under some suitable qualification conditions, the generalized perturbation map \(G\) (that is, the solution set map to a parameterized constraint system, to a parameterized variational inequality, or to a parameterized optimization problem) in \(G(x,z)=\{ y\in D: z\in F(x,y)+K(y)\}\) (where \(F:R^d\times R^n = R^m , K: R^n = R^m)\) is proto-differentiable. The precise proofs of the main results are proposed.