WIP: Feature/mixed space
- Jul 06, 2020
-
-
Lisa Julia Nebel authored- Add Multitypesolvers with Multitypecholmod and Flatblockedconverter
-
Lisa Julia Nebel authored
-
Lisa Julia Nebel authored
-
Lisa Julia Nebel authored
- the basis is now assumed to be a compositeBasis of two powerBases (dune-functions compatible and dune-elasticity compatible) - implemented a geodesicfeassemblerwrapper.hh that transforms the assembled matrix and gradient from the mixedgfeassembler to the combined form so the riemanniantrustregionsolver and the riemannianproximalnewtonsolver can still be used - the mixedriemanniantrsolver can also be used -
Lisa Julia Nebel authored
In mixedlocalgfeadolcstiffness: Use the rank as the ADOL-C tape number when calculating the gradient and the hessian
-
Lisa Julia Nebel authored
In mixedriemanniantrsolver and mixedlocalgfeadolcstiffness: Wrap energy calculation and solve-call in try-catch
-
Lisa Julia Nebel authored
Move film-on-substrate.parset-file, dirichlet-values-py file and surface-shell-parameters-py-file from src to problems folder
-
- Jul 02, 2020
-
-
Lisa Julia Nebel authored
Add first version of the Riemannian Proximal Newton solver, as an alternative to the Trust Region Solver. In each step of these two iterative solvers: For a given iterate x, we try to find a correction that decreases the energy of the nonlinear functional. This correction is calculated using the Taylor expansion around x, resulting in the problem: Hessian(x) * correction = -gradient(x). The correction only causes an energy decrease if the functional can be approximated correctly within a certain radius around x. 1) The trust-region algorithm ensures this using a trust-region. 2) The proximal newton method ensures this by "punishing" large corrections using a regularization factor.
-
- Jun 30, 2020
-
-
Lisa Julia Nebel authored
- read in the deformation file on each process - modify trustregionsolver so it can run in parallel
-