B02 Robust sparse low rank approximation of multi-parametric partial differential equations

This project aims at numerical methods for parameter-dependent elliptic PDEs based on polynomial expansions in the parameter domain that retain their efficiency in challenging cases, such as high-contrast coefficients. To this end, we use a new type of low rank approximation for problems with infinitely many parameters and parameter-dependent approximate inverses provided by the new concept of tensor coefficient hierarchical matrices.