Article
Discrepancy Between Qmc and Rqmc
Authors
Abstract
We introduce a class of functions in d ≥ 3 dimensions which have arbitrary odd superposition effective dimensions between three and d inclusive. We prove that for the integration of any function in this class any Sobol’ points of a fixed length have zero error, whereas Owen’s scrambling of any Sobol’ points of the same length has the same variance of error as simple Monte Carlo meth- ods. Furthermore, for any function in the same class Owen’s scrambling of high- discrepancy points, which consist of d copies of the van der Corput points in base two, gives zero-variance estimates for the integration.
Keywords
Effective dimension, generalized Sobol’ sequences, high dimensional integrals.
Citation
Tezuka, S. (2007). Discrepancy between qmc and rqmc. Uniform Distribution Theory, 2(1), 93–105.
S. Tezuka, “Discrepancy between qmc and rqmc,” Uniform Distribution Theory, vol. 2, no. 1, pp. 93–105, 2007.
Tezuka S. Discrepancy between qmc and rqmc. Uniform Distribution Theory. 2007;2(1):93–105.
Tezuka, S. (2007), ‘Discrepancy between qmc and rqmc’, Uniform Distribution Theory, 2(1), pp. 93–105.
Tezuka, Shu. “Discrepancy Between Qmc and Rqmc.” Uniform Distribution Theory, vol. 2, no. 1, 2007, pp. 93–105.
Tezuka, Shu. “Discrepancy Between Qmc and Rqmc.” Uniform Distribution Theory 2, no. 1 (2007): 93–105.
Export citation
Published by: Engineering Journals


