A Robust Preconditioner for Distributed Optimal Control for Stokes Flow

Dipl.-Ing. Markus Kollmann

April 12, 2011, 1:30 p.m. HS 14

In this talk we present an abstract framework how to find norms for parameter-dependent saddle point problems which lead to robust (i.e.: parameter-independent) estimates of the solution in terms of the data. For the distributed optimal control problem for the Stokes equations we derive explicit formulas for these norms. This leads to a block-diagonal preconditioner for the discretized problem with mesh-independent and robust convergence rates if used in preconditioned Krylov subspace methods. Numerical examples are given which illustrate the theoretical results.