Uniform Distribution Theory
Journal license

Journal

Uniform Distribution Theory


Volume
& Issue

Volume 6, Issue 1


Published
on

December 6, 2010


Pages

57-64


DOI

Article

Discrepancy Between Qmc and Rqmc, Ii


Authors

Shu Tezuka Affiliation:
Faculty of Mathematics Kyushu University 744 Motooka, Nishi-ku, Fukuoka-shi, Fukuoka-ken JAPAN 819-0395


Abstract

There are two types of randomization for (t, m, d)-nets: Owen scrambling and random digital shift. In the previous paper [Uniform Distribution Theory, 2 (2007), 93-105], we introduced a class of functions for which any Sobol’ points have zero integration error, whereas Owen scrambling of Sobol’ points has the same variance of integration error as that of simple Monte Carlo methods. In this paper, by using the same functions as the paper mentioned above, we con- struct an example of functions for which any Sobol’ points have zero integration error, whereas not only Owen scrambling but also random digital shift of Sobol’ points have variance of integration error no smaller than that of simple Monte Carlo methods.


Keywords

Generalized Sobol’ sequences, high dimensional integration, Monte Carlo and.


Citation

Tezuka, S. (2011). Discrepancy between qmc and rqmc, ii. Uniform Distribution Theory, 6(1), 57–64.

Published by: Engineering Journals

Engineering Journals Logo