Tensor methods for parametric PDEs
Event details
| Date | 23.04.2024 |
| Hour | 16:15 › 17:15 |
| Speaker | Dr Damiano Lombardi - Inria Paris |
| Location | |
| Category | Conferences - Seminars |
| Event Language | English |
Tensor methods are a way to build parsimonious discretisations of parametric Partial Differential Equations (PDEs). The talk will be divided in two parts. In the first one, the preconditioning of the multi-linear system arising from the discretisation is discussed. Three different strategies are presented, analysed and compared.
In the second part of the talk, the problem of taking the geometrical variability of domains into account is described. A constructive non-linear method is presented to provide a parametric parsimonious representation of sets of characteristic functions which vary in a neighbourhood of a reference configuration. This can be considered as a first step towards Uncertainty Quantification for PDEs in varying domains.
Practical information
- General public
- Free