Sensitivity analysis for variational inclusions by Wiener-Hopf equation techniques (Q1818333)

The authors extend the sensitivity analysis framework developed for variational inequalities to variational inclusions relying on Wiener-Hopf equations technique. They consider a general variational inclusion : find \(x\in X,\) such that \(0\in Ax+B(g(x)),\) where \(X\) is a real Hilbert space, \(A, g\) are nonlinear operators in \(X\) and \(B\) is a maximal monotone operator. Related to this problem they considered the generalized Wiener-Hopf equation: find \(z\in X,\) such that \(Ag^{-1}J^B_\mu (z)+B_\mu z=0,\) where \(\mu>0\), \(J^B_\mu:=(I+\mu B)^{-1}\) and \(B_\mu:=\frac 1\mu (I-J^B_\mu)\) are the resolvent and the Yosida approximate associated with \(B\). The sensitivity analysis is performed with respect to the parameter \(\lambda\in \Lambda,\) where \(\Lambda\) is an open subset of a Hilbert space \(Y\). The parametric version of the variational inclusion is : \(0\in A(x_\lambda,\lambda)+B(g(x_\lambda,\lambda),\lambda),\) where \(A(.,\lambda), B(.,\lambda):X\times \Lambda\to X\) are given operators. The associated parametric Wiener-Hopf equation is: \(Ag^{-1}J_\mu^{B(.,\lambda)}z_\lambda+(B(.,\lambda)_\mu z_\lambda = 0\). The main result expresses the Lipschitz continuity of the functions \(\lambda\to x(\lambda)\) and \(\lambda\to z(\lambda)\).