For large, nonlinear, and time dependent PDE constrained optimization problems with 3D spatial domain, reduced methods are a viable algorithmic approach. The computation of reduced gradients by adjoint methods requires the storage of 4D data, which can be quite expensive from both a capacity and bandwith point of view.
This project investigates lossy compression schemes for storing the state trajectory, based on hierarchical interpolation in adaptively refined meshes as a general predictor.
Further information is available in the detailed project description.
Phone: +49 (0) 30 84185-170
Fax: +49 (0) 30 84185-107
Konrad Polthier (FU Berlin)
Christoph von Tycowicz (FU Berlin)
DFG Research Center Matheon "Mathematics for key technologies", Project F9