Self-Adaptive Reliable Numerical Treatment of Polymorphic Uncertainty by Hierarchical Tensors
Abstract of the planned project content, taking into account the results from the first project phase
In phase two of the project our emphasis is on the construction of framework for computations and postprocessing. One has to keep in mind that the output in compressed form is a hierarchical tensor and thus should be handled with specialised arithmetic procedures in order to retain the extreme speedup.
In the second Phase we will create a framework for computations and postprocessing based upon the PolyHT model. This framework will encompass, a specifically developed sensitivity analysis for polymorphic uncertainty quantification, efficient models for the computation of level sets of PolyHT surrogates and finally methods for parameter estimation. Along the way, the methods for the construction of the PolyHT model will be further refined.
Essential project goals and objectives
- further developed sensitivity analysis
- efficient level set methods for the propagation of fuzzy uncertainty
- parameter estimation in the presence of polymorphic uncertainty
- refinement of the construction of the PolyHT surrogate
Prof. Grasedyck
Prof. Dr. Lars Grasedyck
RWTH Aachen
Institut für Geometrie und praktische Mathematik (IGPM)
Dieter Moser
Dieter Moser
RWTH Aachen
Institut für Geometrie und praktische Mathematik (IGPM)
Kastian, Steffen, Moser, Dieter et al. "A two-stage surrogate model for Neo-Hookean problems based on adaptive proper orthogonal decomposition and hierarchical tensor approximation." Computer Methods in Applied Mechanics and Engineering 372 (2020): 113368.