## Preliminary list of abstracts

Note: We are using the fantastic MathJax JavaScript library to typeset the mathematics on this web page. You can right-click on a formula to zoom in or to select the rendering engine you like (all under 'Settings'). Under Firefox the default renderer turns out to be MathML (which is quick), but you might prefer the HTML-CSS renderer for more faithful LaTeX rendering. If you encounter any problems with this setup: then please email us!

Hierarchical tensor methods for PDEs with stochastic parameters
Keywords: Hierarchical Tucker; Stochastic PDEs; Low Rank Tensors
Tue, 10:45--11:10
• Grasedyck, Lars (IGPM, RWTH Aachen, Germany)
• Ballani, Jonas (Max Planck Institut für Mathematik in den Naturwissenschaften, Leipzig)
• Kluge, Melanie (IGPM, RWTH Aachen)

We consider the problem to solve a (stochastic) parameter dependent equation $A(\omega)u(\omega) = b(\omega),\qquad \omega\in\Omega$ for systems $A$ governed by partial differential operators that depend on $\omega$. Our aim is to calculate quantities of interest (mean, variance, maximum etc.) of the set of solutions. One way to solve such a problem is by expansion of the system, the right-hand side as well as the solution in independent stochastic variables $\omega_1,\ldots,\omega_p$, and then solve the arising large-scale deterministic problem $A(\omega_1,\ldots,\omega_p)u(\omega_1,\ldots,\omega_p) = b(\omega_1,\ldots,\omega_p).$ An alternative approach is to use (quasi or multilevel) Monte Carlo (MC) methods which require just a simple sampling ($M$ simulations), but these are only useful for certain quantities of interest (e.g. the mean). We will present a new approach based on hierarchical Tucker (HT) representations of tensors. This method is based on standard PDE solvers for deterministic systems. The set of solutions is approximated in a low rank (HT) tensor format that allows for many parameters (thousands).