Discretization error analysis for a radially symmetric harmonic map heat flow problem

Nam Anh Nguyen, Arnold Reusken · 2025

Read Discretization error analysis for a radially symmetric harmonic map heat flow problem by Nam Anh Nguyen, Arnold Reusken (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2503-03525.txt.

Introduction

Nam Anh Nguyen and Arnold Reusken bound the error of a numerical scheme for a heat flow of maps in a 2025 paper. The radially symmetric case of harmonic map heat flow is a laboratory equation behind many analyses of the full problem. Their solver pairs second order finite differences in space with a semi implicit Euler march in time, leaving one linear solve per step.

For smooth solutions, they prove error bounds of optimal order apart from a logarithmic factor. Discrete energy estimates mirror the continuous flow's energy decay, a good sign the scheme respects the physics. The analysis gives practitioners a certified baseline before they attack harder, less symmetric versions of the flow.

Catalog record

CREATORNam Anh Nguyen, Arnold Reusken
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE46 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-03525

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Discretization error analysis for a radially symmetric harmonic map heat flow problem Nam Anh Nguyen, Arnold Reusken Abstract In this paper we study the harmonic map heat flow problem for a radially symmetric case. The corresponding partial dfferential equation plays a key role in many analyses of harmonic map heat flow problems. We consider a basic discretization method for this problem, namely a second order finite difference discretization in space combined with a semi-implicit Euler method in time. The semi-implicit Euler method results in a linear problem in each time step. We restrict to the regime of smooth solutions of the continuous problem and present an error analysis of this discretization method. This results in optimal order discretization error bounds (apart from a logarithmic term). We also present discrete energy estimates that mimic the decrease of the energy of the continuous solution. keywords Harmonic map heat flow problem, discrete stability analysis, discretization error bounds, discrete energy estimates. 1 Introduction Let \Omega\subset\mathbb{R}^{N} , N=2,3 , be a Lipschitz domain and S^{2} the unit sphere in \mathbb{R}^{3} . The harmonic map heat flow (HMHF) problem is as follows. Given an initial condition \mathbf{u}_{0}:\Omega\to S^{2} , determine \mathbf{u}(\cdot,t):\Omega\to S^{2} such that \partial_{t}\mathbf{u}=\Delta\mathbf{u}+\rvert\nabla\mathbf{u}\lvert^{2}\mathbf{u},\quad\mathbf{u}(\cdot,0)=\mathbf{u}_{0},\quad\mathbf{u}(\cdot,t)_{|\partial\Omega}=(\mathbf{u}_{0})_{|\partial\Omega},\quad t\in(0,T]. (1) This problem is obtained as the L^{2} gradient flow of the Dirichlet energy \int_{\Omega}|\nabla\mathbf{v}|^{2}\,dx on vector fields \mathbf{v}:\Omega\to\mathbb{R}^{3} that satisfy a pointwise unit length constraint. Unit length minimizers of this Dirichlet energy are called harmonic maps. The HMHF problem is closely related to the Landau-Lifshitz-Gilbert (LLG) equation. The HMHF equation can be considered as the limit of the LLG equation where the precessional term vanishes and only damping is left [ 26 ] . Harmonic maps, HMHF and LLG equations have numerous applications, for example, in the modeling of ferromagnetic materials or of liquid crystals, cf. e.g. [ 24 , 25 , 28 , 21 ] . We refer to [ 32 , Chapter III, Section 6] and references therein for a further introduction to HMHF. There is extensive mathematical literature in which topics related to well-posedness, weak formulations, regularity, blow-up phenomena and convergence of solutions of the HMHF problem to harmonic maps are studied, cf. e.g [ 31 , 32 , 13 , 19 , 23 , 15 , 12 , 35 ] . Early work on the development and analysis of numerical methods for HMHF or LLG problems is found in [ 8 , 7 , 28 , 9 , 2 , 3 ] . In recent years, there has been a renewed interest in the numerical analysis of methods for this problem class [ 1 , 6 , 17 , 5 ] . In this paper we study the HMHF problem for a specific radially symmetric case. Assume \Omega is the unit disk in \mathbb{R}^{2} and assume the solution \mathbf{u} to be radially symmetric. Using polar coordinates on the disk a special type of solution of ( 1 ) is given by \mathbf{u}(r,\psi,t)=\begin{pmatrix}\cos{\psi}\sin{u(r,t)}\\ \sin{\psi}\sin{u(r,t)}\\ \cos{u(r,t)}\end{pmatrix}, (2) and the partial differential equation in ( 1 ) can be reduced to u_{t}=u_{xx}+\frac{1}{x}u_{x}-\frac{\sin(2u)}{2x^{2}}\quad\text{on}~(0,1]\times[0,T],\\ (3) for the function u(x,t)=u(r,t) on [0,1]\times[0,T] , as is shown in [ 30 , Section 2.1] . The result in [ 11 , Lemma 2.2] proves that if the solution \mathbf{u} to ( 1 ) is of the form ( 2 ), then u solves ( 5 ). Note that due to the structure of the solution \mathbf{u} in ( 2 ) the unit length constraint is satisfied. The partial differential equation in ( 3 ) has a strong nonlinearity due to the term \sin(2u) and has a singular behavior for x\downarrow 0 . The Dirichlet energy in terms of the function \mathbf{u} is given by \displaystyle E(\mathbf{u})=\frac{1}{2}\int_{\Omega}|\nabla\mathbf{u}|^{2}\,\mathrm{d}\mathbf{x}, which by transformation into polar coordinates and using radial symmetry yields E(\mathbf{u})=\pi\int_{0}^{1}\big(u_{x}^{2}+\frac{\sin^{2}u}{x^{2}})x\,dx=:\pi\mathcal{E}(u). (4) Equation ( 3 ) has been used in models of nematic liquid crystals [ 35 ] . More importantly, the radially symmetric case ( 3 ) plays a fundamental role in the analysis of HMHF. In the seminal works [ 12 , 11 ] the authors study finite time singularities of ( 1 ) for the two-dimensional case N=2 . In these studies the radially symmetric case ( 3 ) plays a crucial role. It is shown that for this problem with initial data u_{0} with |u_{0}|\leq\pi a unique global smooth solution exists, whereas for the case with |u_{0}(1)|>\pi the solution blows up in finite time. The latter means that for the solution u the derivative at x=0 becomes arbitrary large: \lim_{x\downarrow 0}|u_{x}(x,t)|\to\infty for t\uparrow T_{\rm crit} . In these analyses of the problem ( 3 ) the maximum principle is an important tool. The work [ 12 ] has motivated further investigations of the blow-up behavior, e.g., [ 10 , 33 ] where infinitely many solutions of ( 3 ) are constructed whose energy is bounded by the initial energy for all times t , but can increase at some t>0 , even for a smooth initial condition. A further related topic is the analysis of the blow up rate. Several theoretical aspects of blow up rates of solutions of ( 3 ) are studied in [ 34 , 4 , 29 ] . In a series of works, Gustafson et al. [ 16 , 18 ] investigate the well-posedness and regularity of the m -equivariant version of ( 3 ). Hocquet [ 22 ] has studied well-posedness and regularity of a stochastic version of ( 3 ). There are only very few papers in which numerical aspects of ( 3 ) are treated. In [ 36 ] an adaptive method is presented, based on finite element discretization in space combined with a stiff ODE solver. The paper [ 20 ] treats a moving mesh ansatz for the discretization of ( 3 ), using finite differences in space in combination with an ODE solver for stiff systems. In both works, the authors are interested in capturing the blow up behavior of the solution with their numerical method. We are not aware of any literature in which for a discretization of ( 3 ), even in the regime of smooth solutions, a rigorous error analysis is presented. Such an error analysis is the main contribution of this paper. We consider a very basic discretization of ( 3 ), namely a second order finite difference discretization in space combined with a semi-implicit Euler method in time. The semi-implicit Euler method results in a linear problem in each time step. We restrict to the regime of smooth solutions of the continuous problem and present an analysis based on the classical “stability plus consistency” framework. Note, however, that in the analysis of the consistency error we also have to bound the linearization error. The analyses of blow up behaviour of the continuous problem ( 3 ) show that there is a critical dependence of stability properties of the problem on the size of the initial condition ( |u_{0}|\leq\pi : globally smooth solution; |u_{0}(1)|>\pi : finite time blow up). In view of this, it is not surprising that the key difficulty in the error analysis of the discrete scheme lies in the derivation of satisfactory stability estimates. Our discrete stability analysis uses M -matrix theory, which in a

Read the complete text