Numerical investigation on the Total Sensitivity Index influence in the solution of stochastic partial differential equations
ABGRALL, Remi
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Institut de Mathématiques de Bordeaux [IMB]
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Institut de Mathématiques de Bordeaux [IMB]
CONGEDO, Pietro Marco
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
GERACI, Gianluca
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
ABGRALL, Remi
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Institut de Mathématiques de Bordeaux [IMB]
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Institut de Mathématiques de Bordeaux [IMB]
CONGEDO, Pietro Marco
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
GERACI, Gianluca
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
< Réduire
Parallel tools for Numerical Algorithms and Resolution of essentially Hyperbolic problems [BACCHUS]
Langue
en
Rapport
Ce document a été publié dans
2012-03-21p. 19
Résumé en anglais
In the present work a criterion on the reduction of the stochastic dimensions has been investigated in the context of the stochastic partial differential equations. Three different types of equations have been analyzed: ...Lire la suite >
In the present work a criterion on the reduction of the stochastic dimensions has been investigated in the context of the stochastic partial differential equations. Three different types of equations have been analyzed: elliptical, parabolic and an hyperbolic. For each equation, both mean and variance have been computed on some scalar output of interest. The complete and the reduced models have been compared in terms of statistical moments. The validity and the efficiency of a criterion based on TSI index has been investigated, and an error correlation has been found between the error on the variance and the TSI, that will be validated in a future work for more complex equations.< Réduire
Origine
Importé de halUnités de recherche