Speaker : Dominik Meidner Abstract: Computing Derivatives in PDE Constrained Optimization via the Adjoint Calculus Many algorithms used to solve optimization problems require the computation of derivatives. In this talk, we present an efficient possibility of calculating derivatives in the case of PDE constrained optimization which relies on the solution of additional linear (adjoint) equations. We demonstrate the approach by means of a Newton iteration for an optimization problem governed by a semilinear elliptic PDE.