Adaptive Galerkin FEM for stochastic forward and inverse problems

Add to your list(s) Download to your calendar using vCal

• Martin Eigel (WIAS-Berlin)
• Thursday 15 February 2018, 12:00-13:00
• Lecture Theatre 2, Strathcona Building.

If you have a question about this talk, please contact Sergey Sergeev.

Stochastic Galerkin methods are among the popular methods for solving PDEs with random data numerically, in particular for engineering applications. Despite the rather involved implementation and computational complexity, a unique advantage is the possibility to derive reliable a posteriori error estimators to control the discretization errors. We review results on the affine parametric case and discuss approaches for technically more difficult lognormal coefficients. In order to cope with the high dimensionality, modern hierarchical tensor formats are employed. Furthermore, the parameter to solution map can serve as a basis for stochastic inverse problems. This is examined in the setting of a a fully adaptive Bayesian inversion with explicit functional representations.

This talk is part of the Optimisation and Numerical Analysis Seminars series.

This talk is included in these lists:

• Lecture Theatre 2, Strathcona Building
• Optimisation and Numerical Analysis Seminars
• School of Mathematics events

Note that ex-directory lists are not shown.