Falk Hante, Michael Hintermüller, Sebastian Pokutta
01.04.2021 − 31.03.2024
We derive theory and algorithms to support 0-1 decision making for PDE-constrained optimization problems. We explore extended formulations of such problems suitable for sequential exact and approximative relaxation to obtain new and powerful stationarity concepts as limits of the corresponding sequence of adjoint-based primal-dual optimality conditions. With this we advance decision making for energy distribution networks from stationary to physical models that accurately capture transient effects e.g. from highly volatile renewable resources.