Anisotropic gradient rearrangement of BV functions and applications

Gloria Paoli, Yabo Yang* · 2026

Read Anisotropic gradient rearrangement of BV functions and applications by Gloria Paoli, Yabo Yang* (2026) 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-2605-15849.txt.

Introduction

Gloria Paoli and Yabo Yang give a symmetry method for the gradient of a bounded-variation function in the anisotropic case. Their work extends a Euclidean result of Amato, Gentile, Nitsch and Trombetti from 2024. The key step splits the regular part of the gradient from its singular part. The main result compares the function with its symmetrized form in L1. As a use, they derive isoperimetric bounds for shape measures tied to twist stiffness.

Catalog record

CREATORGloria Paoli, Yabo Yang*
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE45399
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.15849

Browse the full library · About the Tower of Babel

Full text

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

Anisotropic gradient rearrangement of BV functions and applications Gloria Paoli, Yabo Yang* Abstract In this paper, we introduce a symmetrization technique for the distributional gradient of a function of bounded variation in the anisotropic setting. This generalizes the result obtained in the Euclidean case in [ 9 ] by separating the absolutely continuous part of the anisotropic gradient from its singular part. Our main result is an L 1 L^{1} comparison between the function and its anisotropic symmetrization. Moreover, as an application, we derive isoperimetric inequalities for some geometric functionals related to the torsional rigidity. Keywords: Anisotropy, BV functions, rearrangement of gradient MSC 2020: 26A45, 35A23, 35B45; 35J60 1 Introduction Symmetrization and rearrangement methods are a classical tool to compare solutions of PDEs defined on a general domain \Omega\subset\mathbb{R}^{n} with suitably symmetric solutions defined on a ball (or, more generally, on an isoperimetric minimizer) having the same volume. Starting from the pioneering work of Talenti [ 35 ] , such techniques yield sharp comparison principles and extremal inequalities for elliptic problems, and they are tightly connected with isoperimetric-type inequalities. The references for the study of the symmetrization of the Laplace operator is very wide, see e.g. [ 1 , 28 , 29 , 36 , 37 , 38 ] for the Dirichlet boundary conditions, [ 4 , 14 , 31 , 32 ] for Robin boundary conditions, and [ 17 , 18 , 34 ] for the anisotropic setting. The starting point of this work is the following comparison result, concerning solutions to Hamilton-Jacobi equations with Dirichlet boundary conditions, proved in [ 27 ] . Theorem 1.1 (Giarrusso-Nunziante) . Let p\geq 1 , f:\Omega\rightarrow\mathbb{R} , F:\mathbb{R}^{n}\rightarrow\mathbb{R} be measurable non-negative functions and let K:[0,+\infty)\rightarrow[0,+\infty) be a strictly increasing function such that 0\leq K(|y|)\leq F(y),\quad\forall y\in\mathbb{R}^{n},\quad\quad K^{-1}(f)\in L^{p}(\Omega) Let u\in W^{1,p}_{0}(\Omega) be a function that satisfies \displaystyle\begin{cases}F(\nabla u)=f(x)&\text{in }\Omega,\\ u=0&\text{on }\partial\Omega,\end{cases} (1) and let v be the unique decreasing spherically symmetric solution to \displaystyle\begin{cases}K(|\nabla v|)=f_{\#}(x)&\text{in }\Omega^{\#},\\ v=0&\text{on }\partial\Omega^{\sharp},\end{cases} (2) where \Omega^{\sharp} is the ball centered at the origin with the same measure as \Omega , and f_{\#} is the increasing rearrangement of f (see Section 2 for the definition). Then, we have the following \|u\|_{L^{1}(\Omega)}\leq\|v\|_{L^{1}(\Omega^{\sharp})}. We refer to [ 12 , 33 ] for the rigidity results, and to [ 5 , 21 , 25 ] for L^{q} comparison. Moreover, we refer to [ 10 ] for the comparison of solution to Hamilton-Jacobi equations with non-zero boundary trace, and to [ 8 ] for quantitative type results. A more delicate question concerns rearrangements of the gradient itself, and the extension of these ideas to the space BV , where the distributional derivative splits into an absolutely continuous part and a singular part. Sharp results in this direction go back to [ 19 ] , where the interplay between Schwarz symmetrization and the structure of Du for u\in\text{BV}(\mathbb{R}^{n}) is studied, proving a Pólya-Szegö inequality for BV functions. Recently, in [ 9 ] , the authors introduced a new symmetrization procedure for BV functions which preserves the total variation while separating the absolutely continuous and singular parts of the derivative: the absolutely continuous part is rearranged in a precise way, and the singular part is concentrated on the boundary of the symmetrized domain. The main result in [ 9 ] is an L^{1} -comparison between u and its gradient symmetrization. The purpose of this paper is to extend the construction and the comparison principle of [ 9 ] to the anisotropic setting. Anisotropic surface energies naturally arise in several models (e.g. crystalline surface tension and interfacial energies), where the cost of a hypersurface depends on the direction of its normal. Anisotropy is encoded by a convex, positively 1 -homogeneous function H:\mathbb{R}^{n}\to[0,+\infty) , with polar function H^{\circ} . In this framework, the Euclidean perimeter is replaced by the anisotropic perimeter P_{H}(E)=\int_{\partial^{*}E}H(\nu_{E})\,d\mathcal{H}^{n-1}, and the role of the Euclidean ball is played by the Wulff ball, which minimizes P_{H} at fixed volume. Our main result provides an anisotropic counterpart of the gradient symmetrization in [ 9 ] . Given a non-negative function u\in\text{BV}_{0}(\Omega) , we construct a function u^{\star} on the Wulff ball \Omega^{\star} with the same volume of \Omega , in such a way that: • the distribution of the anisotropic absolutely continuous energy density H(\nabla^{a}u) matches that of H(\nabla u^{\star}) , namely \big(H(\nabla^{a}u)\big)^{*}=\big(H(\nabla u^{\star})\big)^{*}; here f^{*} denotes the decreasing rearrangement of a measurable function f (see Section 2.3 ); • the singular part of the distributional derivative Du is not lost under symmetrization: its contribution is encoded into the boundary trace of u^{\star} on \partial\Omega^{\star} , so that D^{s}u appears as an explicit boundary term in the comparison problem for u^{\star} ; • and the following L^{1} comparison holds: \|u\|_{L^{1}(\Omega)}\leq\|u^{\star}\|_{L^{1}(\Omega^{\star})}. For the exact definition of u^{\star} see ( 27 ). In particular, this procedure not only symmetrizes the absolutely continuous energy, but also keeps track of D^{s}u through a boundary datum on \partial\Omega^{\star} , making the BV nature of the argument fully explicit. Therefore, our main theorem can be stated as follows. We use the following notation: BV_{0}(\Omega)=\{u\in\text{BV}(\mathbb{R}^{n}):u\equiv 0\,\,\mathrm{in}\,\,\mathbb{R}^{n}\backslash\Omega\}. Theorem 1.2 . Let \Omega\subset\mathbb{R}^{n} be an open bounded set with finite perimeter and let \Omega^{\star} be the Wulff ball centered at the origin with |\Omega^{\star}|=|\Omega| . Assume that u is a non-negative function in \text{BV}_{0}(\Omega) and let u^{\star} be defined by ( 27 ). Then \|u\|_{L^{1}(\Omega)}\leq\|u^{\star}\|_{L^{1}(\Omega^{\star})}. (3) Finally, we illustrate how this anisotropic symmetrization principle can be employed in comparison arguments for variational problems and PDEs, where the natural geometry is governed by H and the extremal domain is a Wulff shape. In particular, functionals involving anisotropic Dirichlet energies and boundary contributions can be handled by combining the main theorem with the interpretation of boundary terms through the singular part of the distributional derivative of suitable zero extensions. Plan of the paper. In Section 2 we recall basic facts on the anisotropic setting, BV functions and rearrengements. In Section 3 we define the anisotropic gradient symmetrization and prove the L^{1} comparison theorem. In the final section, we discuss applications to variational problems in the anisotropic framework. 2 Preliminaries In this section, we fix the notation and recall some basic definitions and properties concerning anisotropic norms (also known as Finsler norms), functions of bounded variation in the anisotropic setting, and rearrangements. We first introduce some basic notations. We will denote by \chi_{\Omega} the characteristic of a set \Omega and by |\Omega| the measure of \Omega . The perimeter of a set \Omega defined as P(\Omega) . Here B_{R}(x) denotes the closed ball of radius R centered at x and B_{R} centered at origin. 2.1 Anisotropic norm and Wulff shape For this section we refer to [ 1 , 16 ] . Throughout the paper we assume that H:\mathbb{R}^{n}\to[0,+\infty) is a norm such that H\in C^{2}(\mathbb{R}^{n}\setminus\{0\}), and such that the unit ball of H is strictly convex and has C^{2} boundary. The function H is called anisotropic norm or Finsler norm, and, often, it represents the surface tension density. In particular, H(\xi)>0\quad\text{for }\xi\neq 0,\qquad H(t\xi)=|t|H(\xi)\quad\text{for every }t\in\mathbb{R},\ \xi\in\mathbb{R}^{n}. (4) Since H is a norm, we have that H(-\xi)=H(\xi)\qquad\text{for every }\xi\in\mathbb{R}^{n}. Moreover, we have that the anisotropy H satisfies \alpha|\xi|\leq H(\xi)\leq\beta|\xi|, (5) for some positive constants \alpha\leq\beta .

Read the complete text