Weighted Sobolev Inequalities via the Meyers–Ziemer Framework: Measures, Isoperimetric Inequalities, and Endpoint Estimates
Simon Bortz, Kabe Moen, Andrea Olivo, Carlos Pérez, Ezequiel Rela · 2026
Read Weighted Sobolev Inequalities via the Meyers–Ziemer Framework: Measures, Isoperimetric Inequalities, and Endpoint Estimates by Simon Bortz, Kabe Moen, Andrea Olivo, Carlos Pérez, Ezequiel Rela (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-2603-05382.txt.
Introduction
Simon Bortz and team set up a global endpoint Sobolev inequality stretching Meyers-Ziemer with a right-side maximal function. The result stretches to weighted bounded change with capacity inequalities. It connects to fractional-operator endpoint figures, including maximal function and Riesz likely bounds. Their main inequality yields endpoint inequalities through subrepresentation formulas and isoperimetric inequalities. Away from the endpoint, analogous inequalities fail, but a bumped maximal function checks them. The method yields novel (p,p) two-weight Sobolev inequalities.
Catalog record
| CREATOR | Simon Bortz, Kabe Moen, Andrea Olivo, Carlos Pérez, Ezequiel Rela |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 114379 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.05382
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Weighted Sobolev Inequalities via the Meyers–Ziemer Framework: Measures, Isoperimetric Inequalities, and Endpoint Estimates Thanks: S. Bortz was supported by the Simons Foundation’s Travel Support for Mathematicians program and an AWI- CONSERVE Fellowship. Thanks: A. Olivo is supported by grants PID2023-146646NB-I00 funded by MICIU/AEI/10.13039/501100011033 and by ESF+, the Basque Government through the BERC 2022–2025 program, and through BCAM Severo Ochoa accreditation CEX2021-001142-S / MICIN / AEI / 10.13039/501100011033. Thanks: C. Pérez is supported by the Spanish government through the grant PID2023-146646NB-I00 and by Severo Ochoa accreditation CEX2021-001142-S both at BCAM and also by the Basque Government through grant IT1615-22 at the University of the Basque Country and by the BERC programme 2022-2025 at BCAM Simon Bortz, Kabe Moen, Andrea Olivo, Carlos Pérez, Ezequiel Rela Abstract . We establish a new global endpoint Sobolev inequality for measures that extends the classical theorem of Meyers–Ziemer by placing a maximal function on the right-hand side. This result has several significant consequences. It extends naturally to functions of weighted bounded variation and yields corresponding capacity and isoperimetric inequalities. The inequality is also closely connected to endpoint estimates for fractional operators, including bounds for fractional maximal functions and Hardy space endpoint estimates for the Riesz potential. Our main inequality yields a family of endpoint inequalities, characterized in terms of subrepresentation formulas, Lorentz space improvements, and isoperimetric inequalities for measures and bounded open sets. When one moves away from the endpoint to p > 1 p>1 , the analogous inequalities no longer hold in general; however, we identify a sharp bumped maximal function for which the corresponding non-endpoint inequality is valid. Finally, we show that this framework yields new ( p , p ) (p,p) two-weight Sobolev inequalities. 2020 Mathematics Subject Classification Primary 46E35; Secondary 26D10, 42B25, 49Q20 1. Introduction 1.1. Classical Sobolev and Isoperimetric Inequalities Consider the classical Gagliardo-Nirenberg-Sobolev (GNS) inequality: \|u\|_{\textup{L}^{p^{*}}(\mathbb{R}^{n})}\leq s_{n,p}\|\nabla u\|_{\textup{L}^{p}(\mathbb{R}^{n})}, (1) which holds for 1\leq p<n , p^{*}=\frac{np}{n-p} , and u\in\textup{C}_{c}^{\infty}(\mathbb{R}^{n}) . If one takes \dot{\textup{W}}^{1,p}(\mathbb{R}^{n}) to be the completion of \textup{C}_{c}^{\infty}(\mathbb{R}^{n}) with respect to the norm \|\nabla u\|_{\textup{L}^{p}(\mathbb{R}^{n})} (which is not a norm in general, but it is a norm on \textup{C}_{c}^{\infty}(\mathbb{R}^{n}) ), then one obtains the Sobolev embedding \textup{W}^{1,p}(\mathbb{R}^{n})\hookrightarrow\textup{L}^{p^{*}}(\mathbb{R}^{n}). The endpoint GNS inequality, \|u\|_{\textup{L}^{\frac{n}{n-1}}(\mathbb{R}^{n})}\leq s_{n,1}\|\nabla u\|_{\textup{L}^{1}(\mathbb{R}^{n})}, (2) is particularly important. First, it implies the general case 1<p<n by scaling the function. More importantly, it is equivalent to the isoperimetric inequality: \Per(\Omega)\geq c_{n}\mathcal{L}(\Omega)^{\frac{n-1}{n}}, where \Per(\Omega) denotes the perimeter of the set \Omega and \mathcal{L} is Lebesgue measure on \mathbb{R}^{n} . For sufficiently regular sets, the perimeter coincides with the Hausdorff measure of the boundary \partial\Omega . The sharp constant in the Sobolev inequality was found by Talenti [ 58 ] and gives the best constant in the isoperimetric equality via c_{n}=s_{n,1}^{-1}=nv_{n}^{1/n}, where v_{n}=\mathcal{L}(B_{1}(0))=\frac{\pi^{n/2}}{\Gamma\big(\frac{n}{2}+1\big)}, is the volume of the unit ball in \mathbb{R}^{n} . Estimate ( 2 ) was extended in [ 41 ] to the setting of locally finite Borel measures \mu as follows: there exists a dimensional constant c_{n} such that, for any locally finite Borel measure \mu and any Lipschitz function u with compact support, \left(\int_{\mathbb{R}^{n}}|u(x)|^{n^{\prime}}\,{\mathrm{d}\mu}(x)\right)^{\frac{1}{n^{\prime}}}\leq c_{n}\int_{\mathbb{R}^{n}}|\nabla u(x)|(\textup{M}\mu(x))^{\frac{1}{n^{\prime}}}\,{\mathrm{d}x}, (3) where M denotes the usual Hardy–Littlewood maximal operator and n^{\prime}=\frac{n}{n-1}=1^{*} . As a consequence of this estimate, an isoperimetric inequality involving measures was also established in [ 41 ] : if \Omega\subseteq\mathbb{R}^{n} is a bounded domain with smooth boundary \partial\Omega and \mu is a measure, then \mu(\Omega)^{\frac{1}{n^{\prime}}}\leq c_{n}\int_{\partial\Omega}(\textup{M}\mu)^{\frac{1}{n^{\prime}}}\,\mathrm{d}\mathcal{H}^{n-1}, (4) where \mathcal{H}^{n-1} denotes the (n-1) -dimensional Hausdorff measure (see Section 3 ). The main objectives of this paper are to show that ( 3 ) is in fact part of a broader family of endpoint measure-theoretic Sobolev inequalities, and to extend the isoperimetric inequality ( 4 ) to more general sets by substantially relaxing the smoothness assumption on the boundary. We emphasize that the constant c_{n} above is not necessarily the same as in ( 3 ). It would be interesting to determine whether the sharp constant in this setting agrees with the classical isoperimetric constant, namely in the case \mu=\mathcal{L} . 1.2. Weighted and Two-Weight Sobolev Inequalities We are interested in GNS inequalities under a change of measure. As motivation, consider a quasiconformal change of variables in inequality ( 1 ). Let f:\mathbb{R}^{n}\to\mathbb{R}^{n} be a bijective map in \textup{W}_{\textup{loc}}^{1,1}(\mathbb{R}^{n},\mathbb{R}^{n}) satisfying |Df(x)|^{n}\leq KJ_{f}(x), where Df(x) is the Jacobian matrix, |Df(x)| its operator norm, and J_{f}(x)=\det Df(x) the Jacobian determinant. After the change of variables inequality ( 1 ) becomes \left(\int_{\mathbb{R}^{n}}|u(x)|^{p^{*}}J_{f}(x)\,{\mathrm{d}x}\right)^{\frac{1}{p^{*}}}\leq s_{n,p}K^{\frac{1}{n}}\left(\int_{\mathbb{R}^{n}}|\nabla u(x)|^{p}J_{f}(x)^{1-\frac{p}{n}}\,{\mathrm{d}x}\right)^{\frac{1}{p}},\qquad u\in\textup{C}_{c}^{\infty}(\mathbb{R}^{n}). This leads naturally to the study of general weighted Sobolev inequalities of the form \|u\|_{\textup{L}^{p^{*}}(w)}\leq C\|\nabla u\|_{\textup{L}^{p}(w^{1-\frac{p}{n}})},\qquad 1\leq p<n. (5) In general, the precise conditions on a weight w under which ( 5 ) holds are difficult to characterize. One approach, introduced by David and Semmes [ 14 ] (see also Semmes [ 52 ] ), is based on the notion of strong A_{\infty} weights , also referred to as metric doubling measures . We will return to the work of David and Semmes in Section 6 . An alternative approach to that of David and Semmes is the following. Consider a weight w such that, for some constant c_{w} , the following estimate holds for any Lipschitz function u with compact support: \|u\|_{\textup{L}^{n^{\prime},\infty}(w)}\leq c_{w}\|\nabla u\|_{\textup{L}^{1}(w^{\frac{1}{n^{\prime}}})}. (6) Then, by the truncation method [ 32 ] (see Theorem 5.1 in Section 5 ), there exists a universal constant \lambda such that we can lift this to the strong estimate ( 5 ), namely: \|u\|_{\textup{L}^{n^{\prime}}(w)}\leq c_{w}\lambda\|\nabla u\|_{\textup{L}^{1}(w^{\frac{1}{n^{\prime}}})}. Now, using a scaling argument (Theorem 6.1 in Section 6 ), we obtain \|u\|_{\textup{L}^{p^{*}}(w)}\leq c_{w,p}\|\nabla u\|_{\textup{L}^{p}(w^{1-\frac{p}{n}})},\qquad 1\leq p<n. (7) Thus, the fundamental issue is to identify the minimal condition required on w to derive the weak type estimate ( 6 ). This is formulated as Conjecture 6.2 in Section 6 . Now, in view of ( 3 ), it is natural to ask whether an analogue of ( 3 ) remains valid for the exponent pair (p,p^{*}) with p>1 . In other words, can one establish an \textup{L}^{p} version of ( 3 ), namely \left(\int_{\mathbb{R}^{n}}|u(x)|^{p^{*}}w(x)\,{\mathrm{d}x})\right)^{\frac{1}{p^{*}}}\leq c_{n,p}\left(\int_{\mathbb{R}^{n}}|\nabla u(x)|^{p}\textup{M}w(x)^{1-\frac{p}{n}}\,{\mathrm{d}x}\right)^{\frac{1}{p}}\text{?} (8) In fact, this estimate is false. One must instead replace the operator M by larger iterated operators \textup{M}^{k} , where the integer k depends on p (see Theorem 2.16 ). We refer the reader to Section 5 for further discussion and related results. On the other hand, it is possible to deduce the following estimate by using ( 3 ) and a scaling variant of the argument described above (see the proof of Theorem 6.1 ). The resulting inequality applies to weights rather than general measures. Specifically, for any weight w on \mathbb{R}^{n} and 1\leq p<n , we have \left(\int_{\mathbb{R}^{n}}|u(x)|^{p^{*}}\,w(x)\,{\mathrm{d}x}\right)^{\frac{1}{p^{*}}}\leq c_{n,p}\left(\int_{\mathbb{R}^{n}}|\nabla u(x)|^{p}\frac{\textup{M}w(x)^{\frac{p}{n^{\prime}}}}{w(x)^{p-1}}\,{\mathrm{d}x}\right)^{\frac{1}{p}}. (9) Observe that the weight on the right-hand side is pointwise larger than w^{1-\frac{p}{n}} appearing in ( 7 ). Inequality ( 9 ) follows as a corollary of [ 40 , Theorem 1.21] , where a localized version was established using a localized form of ( 3 ). To prove inequality ( 6 ), we can invoke the classical theory of Muckenhoupt weights via the Riesz potential operator \textup{I}_{\alpha}f(x)=\frac{1}{\gamma(\alpha)}\int_{\mathbb{R}^{n}}\frac{f(y)}{|x-y|^{n-\alpha}}\,{\mathrm{d}y},\qquad 0<\alpha<n, where \gamma(\alpha)=\frac{2^{\alpha}\pi^{n/2}\Gamma(\alpha/2)}{\Gamma((n-\alpha)/2)}. The Riesz potential \textup{I}_{1} plays a central role in Sobolev inequalities due to the subrepresentation formula |u(x)|\leq\frac{\gamma(1)}{\omega_{n-1}}\,\textup{I}_{1}(|\nabla u|)(x),\qquad u\in\Lip_{c}(\mathbb{R}^{n}), (10) where \omega_{n-1}=nv_{n} denotes the surface measure of the unit sphere. Although