Numerical solution of the Hartree-Fock equation in multilevel tensor-structured format

revised version: November 2009
Boris N. Khoromskij, Venera Khoromskaia, and Heinz-Jürgen Flad
(Please use for correspondence this email).

Submission date: 24. Jul. 2009
Pages: 26
MSC-Numbers: 65F30, 65F50, 65N35, 65F10
Keywords and phrases: Hartree-Fock equation, Tucker/canonical models, discrete multivariate convolution, Tensor-truncated methods, multilevel SCF iteration
Download preprint: PDF (477 kB)

Abstract:
In this paper, we describe a novel method for robust and accurate iterative solution of the self-consistent Hartree-Fock equation in based on the idea of tensor-structured computation of the electron density and the nonlinear Hartree and (nonlocal) Hartree-Fock exchange operators at all steps of the iterative process. We apply the self-consistent field (SCF) iteration to the Galerkin discretisation in a set of low separation rank basis functions that are solely specified by the respective values on the 3D Cartesian grid. The approximation error is estimated by , where is the mesh size of tensor grid, while the numerical complexity to compute the Galerkin matrices scales linearly in n. We propose the tensor-truncated version of the SCF iteration using the traditional direct inversion in the iterative subspace (DIIS) scheme enhanced by the multilevel acceleration with the grid dependent termination criteria at each discretization level. This implies that the overall computational cost scales linearly in the univariate problem size n. Various numerical illustrations are presented for the all electron case of HO, and pseudopotential case of CH and CHOH molecules. The proposed scheme is not restricted to a priori given analytically integrable and/or rank-1 basis sets, that opens further perspectives for promotion of the tensor-structured methods in computational quantum chemistry.