abstract:
We discuss a stochastic model which yields a class of random functions $X$ on a bounded
domain $\mathcal{O}\subset\mathbb{R}^d$ in terms of wavelet expansions. We use these processes because they permit
explicit control of their Besov regularity.
In this talk we present error bounds for linear and non-linear approximation of $X$.
As an outlook we consider $X$ to be the right hand side of an elliptic boundary value problem
an present error bounds for best N-term approximation of the solution $U$.
This is joint work with F. Lindner, R. Schilling (Dresden), K. Ritter (Kaiserslautern), T. Raasch
(Mainz), P. Cioica, S. Dahlke and S. Kinzel (Marburg). |