Premium
Trivial reductions of dimensionality in the propagation of uncertainties: a physical example
Author(s) -
Bolado Ricardo,
Mira José
Publication year - 2004
Publication title -
environmetrics
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.68
H-Index - 58
eISSN - 1099-095X
pISSN - 1180-4009
DOI - 10.1002/env.626
Subject(s) - variance (accounting) , estimator , mathematics , variance reduction , variable (mathematics) , dimensionality reduction , distribution (mathematics) , statistics , reduction (mathematics) , computer science , mathematical analysis , monte carlo method , geometry , accounting , artificial intelligence , business
When performing uncertainty analysis on a mathematical model of a physical process, some coefficients of the differential equations appear as a result of elementary operations of other coefficients. It is shown in this article that variance reduction techniques should be applied on the ‘final’ or ‘reduced’ coefficients and not on the original ones, thus reducing the variance of the estimators of the parameters of the output variable distribution. We illustrate the methodology with an application to a physical problem, a radioactive contaminant transport code. A substantial variance reduction is achieved for the estimators of the distribution function, the mean and the variance of the output. Copyright © 2003 John Wiley & Sons, Ltd.