Lecture 2: Complexity results for integration.

Presented by:
Erich Novak Friedrich-Schiller-Universität Jena
Date:
Wednesday 13th February 2019 - 15:00 to 16:30
Venue:
INI Seminar Room 1
Abstract:
We give a short introduction to IBC and present some basic definitions and a few results. The general question is: How many function values (or values of other functionals) of $f$ do we need to compute $S(f)$
up to an error $\epsilon$? Here $S(f)$ could be the integral or the maximum of $f$.
In particular we study the question: Which problems are tractable? When do we have the curse of dimension?
In this second talk we discuss complexity results for numerical integration. In particular we present results for the star discrepancy, the curse of dimension for $C^k$ functions, and results for randomized algorithms

The video for this talk should appear here if JavaScript is enabled.
If it doesn't, something may have gone wrong with our embedded player.
We'll get it fixed as soon as possible.
Presentation Material: