Core Concepts
Developing a low-rank tensor product framework for radiative transfer equations in plane-parallel geometry.
Abstract
The content discusses the challenges of classical numerical algorithms in solving radiative transfer problems and proposes a low-rank tensor product framework to address dimensionality issues. It introduces a preconditioned and rank-controlled Richardson iteration method, leveraging exponential sums approximations for efficient computation. The paper outlines the weak formulation, Galerkin approximation, and matrix representations involved in the process. Additionally, it explores the construction of suitable preconditioners using Kronecker sum structures and presents numerical experiments to validate the proposed approach.
Stats
J ∼h^-d and N ∼h^(-d+1)
Storage of approximate solution proportional to O(h^-2d+1)