Detail
Random Bits vs. Random Numbers for Numerical Integration
Lecture: Prof. Dr. Klaus Ritter, TU Kaiserslautern
Abstract: Numerical integration is an important application area for randomized (or Monte Carlo) algorithms. In this talk we compare two classes of such algorithms that have access to a generator either for random numbers from [0,1] or for random bits only.
We discuss classical results for finite-dimensional integration w.r.t. the uniform dis¬tribution on [0,1]d as well as recent results for infinite-dimensional integration, which naturally arises, e.g., in the context of stochastic differential equations. The study of the latter problem is partially motivated by computations on FPGAs (field programmable gate array) with applications in mathematical finance.
This talk is based on joint work with M. Giles (Oxford), M. Hefter, and L. Mayer (both from Kaiserslautern).
Additional Information
Open to | all |
Prior registration | not required |
Organised by | Fakultät für Informatik und Mathematik, Lst. Prof. Dr. T. Müller-Gronbach |
Contact organizer of event | alexandra.vogl@uni-passau.de |