Stochastic Galerkin method for elliptic SPDEs: a white noise approach (Q2471407)

This paper considers the use of stochastic Galerkin methods for elliptic SPDEs, that arise as models in many applications in the natural sciences. The diffusivity is considered as a random field, a fact which turns the solution of the equation into random fields as well. The authors use a white noise framework in order to treat this problem, and employ the Wiener-Itô chaos expansion to provide a spectral decomposition of the diffusivity, with special emphasis placed on the constraints of positivity of the diffusion coefficient, needed to guarantee the elipticity of the SPDE. After introducting the proper functional analytic setup, the authors give variational and Galerkin formulations of the problem and using versions of the Lax-Milgram Lemma argue concerning existence, uniqueness and stability results. The test function \(\phi\) (which belongs to the Schwartz space of rapidly decreasing functions) which is used in the Bochner-Minlos theorem for the construction of the Gaussian measure plays an important role in the method. The final part of the paper discusses the advantages of using the white noise framework in the spectral approach, as well as the effect of the choice of the test function \(\phi\) on the covariance function for the solution.