Logo Leibniz Universität Hannover
Logo: Institut für Baumechanik und Numerische Mechanik/Leibniz Universität Hannover
Logo Leibniz Universität Hannover
Logo: Institut für Baumechanik und Numerische Mechanik/Leibniz Universität Hannover
  • Zielgruppen
  • Suche
 

Scientific presentation about "Sparse Grid Collocation" :: 22.06.2018 :: 2 pm

In this talk Dr. Björn Sprungk (Institute for Mathematical Stochastics, Georg-August-Universität Göttingen) considers the problem of approximating a computationally expensive function — such as the solution of a parametric differential equation — depending on a possibly large number of variates.

This problem arises particularly in uncertainty quantification (UQ) with complex computer models where the variates represent uncertain model parameters and the approximation of the complex model shall reduce the computational effort of the UQ procedure. Dr. Björn Sprungk focuses on multivariate polynomial interpolation and discusses the sparse grid approach as a way to break the curse of dimensionality, i.e., the computational effort does not grow exponentially w.r.t. the dimension, in fact it grows only polynomially.

The idea of sparse grids dates back to Russian mathematicians in the 1970s and they were again promoted for numerical simulation by Christoph Zenger in the early 1990s.

  • Speaker: Dr. Björn Sprungk, Institute for Mathematical Stochastics, Georg-August-Universität Göttingen
  • Venue: MZ2 (010) – Hochhaus Appelstraße 9a

Keywords: uncertainty quantification, stochastic collocation, sparse grids, groundwater flow simulation

Further information

We are pleased to host Dr. Björn Sprungk presenting his long-term experience in stochastic sparse grid col- location. All those who are interested in this topic are kindly invited to join his talk and afterwards discussion.