scientific article; zbMATH DE number 1136325

From MaRDI portal
Publication:4381501





\textit{E.-B. Lin} and \textit{X. Zhou} [Numer. Methods Partial Differ. Equations 13, No. 4, 303-320 (1997; Zbl 0881.65097)] gave a wavelet approximation theorem under rather strict conditions for the mother wavelet and the corresponding scaling function. They used sampling values \(f_{kl}^j=\frac{1}{2^j}f(\frac{k+\lambda }{2^j},\frac{l+\lambda}{2^j}).\) Here the authors introduce multi-scaling function interpolation (MSF interpolation) by choosing the sample values with suitable weighted average of the values of a given function at the sampling points. They show that the necessary and sufficient condition for the asymptotic behaviour of MSF interpolation with error \(2^{-jN}\) in the \(L_2\) space or \(2^{-j(N-n)}\) in the Sobolev space \(H^n\) is that the multi-scaling function \(\Phi (t)=[\Phi _0(t),\ldots ,\Phi_{r-1}(t)]^T\) has approximation accuracy \(N\) (i.e., each polynomial \(x^p,p=0,\ldots,N-1,\) is a locally finite series of the integer translates \(\Phi (x+k)\), \(k\in \mathbb Z\)). Numerical examples and several special cases are presented, too. For the entire collection see [Zbl 0884.00041].











This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4381501)