High-order numerical discretizations and a posteriori error estimates for variational inequalities (Jad Dabaghi, Sorbonne Université)

We propose an adaptive inexact version of a class of semismooth Newton methods for variational inequalities. As a model problem, we study the system of variational inequalities describing the contact between two membranes. We study a family of Galerkin numerical schemes that discretize this problem. We consider any iterative semismooth linearization algorithm like the Newton-min or the Newton–Fischer–Burmeister which we complement by any iterative linear algebraic solver. In the case of finite elements, we then derive an a posteriori estimate on the error between the exact solution at the continuous level and the approximate solution which is valid at any step of the linearization and algebraic resolutions. Our estimate is based on flux reconstructions in discrete subspaces of H(div,Ω) and on potential reconstructions in discrete subspaces of H1 (Ω) satisfying the constraints. It distinguishes the discretization, linearization, and algebraic components of the error. Consequently, we can formulate adaptive stopping criteria for both solvers, giving rise to an adaptive version of the considered inexact semismooth Newton algorithm. Under these criteria, the efficiency of the leading estimates is also established, meaning that we prove them equivalent with the error up to a generic constant. Numerical experiments for the Newton-min algorithm in combination with the GMRES algebraic solver confirm the efficiency of the developed adaptive method. An extension to unsteady problems is also discussed in the present work.