Novel results for the anisotropic sparse quadrature and their impact on random diffusion problems

This article is dedicated to the anisotropic sparse Gaussian quadrature for functions which are analytically extendable into an anisotropic tensor product domain. Based on a novel estimate for the cardinality of the anisotropic index set, we are able to substantially improve the error versus cost estimates of the anisotropic sparse quadrature. To validate the theoretical findings, we use the anisotropic sparse Gaussian quadrature to compute the moments of elliptic partial differential equations with random diffusion.