NANAJul 3, 2018

Maximum norm error estimates for the finite element approximation of parabolic problems on smooth domains

arXiv:1805.013363 citationsh-index: 16
Originality Synthesis-oriented
AI Analysis

Provides theoretical error bounds for finite element methods on non-convex smooth domains, addressing a known gap in the literature.

This paper derives $L^\\infty$-error estimates for finite element approximation of parabolic problems on smooth domains with inhomogeneous Neumann boundary conditions, showing that the convergence rate is suboptimal for higher-order elements due to the non-empty symmetric difference between the approximate and original domains.

In this paper, we consider the finite element approximation for a parabolic problem on a smooth domain $Ω\subset \mathbb{R}^N$ with the inhomogeneous Neumann boundary condition. We emphasize that the domain can be non-convex in general. We implement the finite element method for this problem by constructing a family of polygonal or polyhedral domains $\{ Ω_h \}_h$ that approximate the original domain $Ω$. The main result of this study is the $L^\infty$-error estimate for this approximation. We shall show that the convergence rate is not optimal for higher order elements since the symmetric difference $Ω\bigtriangleup Ω_h$ is not empty in general. In order to address the effect of the symmetric difference of domains, we introduce the tubular neighborhood of the original boundary $\partialΩ$. We will also present a slightly new approach to establish the $L^\infty$-error estimate. Moreover, we present the smoothing property for the discrete parabolic semigroup and the spatially discretized maximal regularity as corollaries of the main result.

Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes