Approximation of Functions from a Hilbert Space Using Function Values or General Linear Information

  • Ralph Tandetzky
Conference paper
Part of the Springer Proceedings in Mathematics & Statistics book series (PROMS, volume 23)


We give an example of a Hilbert space embedding \(H \subset {\mathcal{l}}_{p}\), \(1 \leq p < \infty \), whose approximation numbers tend to zero much faster than its sampling numbers. The main result shows that optimal algorithms for approximation that use only function evaluation can be more than polynomially worse than algorithms using general linear information.



I want to thank Erich Novak and Aicke Hinrichs for valuable discussions and ideas which lead to the results in this paper.


  1. 1.
    A. Hinrichs, E. Novak, J. Vybíral: Linear information versus function evaluations for L 2-approximation, J. Approx. Theory 152 (2008), pp. 97–107.Google Scholar
  2. 2.
    F. Y. Kuo, G. W. Wasilkowski, H. Woźniakowski: On the power of standard information for multivariate approximation in the worst case setting, J. Approx. Theory 158 (2009), pp. 97–125.Google Scholar
  3. 3.
    E. Novak, H. Woźniakowski: On the power of function values for the approximation problem in various settings.Google Scholar
  4. 4.
    J. F. Traub, H. Woźniakowski: A General Theory of Optimal Algorithms, Academic Press, New York, 1980.Google Scholar
  5. 5.
    G. W. Wasilkowski, H. Woźniakowski: On the power of standard information for weighted approximation, Found. Comput. Math. 1 (2001), pp. 417–434.Google Scholar

Copyright information

© Springer-Verlag Berlin Heidelberg 2012

Authors and Affiliations

  1. 1.Mathematisches InstitutUniversity of JenaJenaGermany

Personalised recommendations