Barycentric stability of nonlocal perimeters: the convex case
Chiara Gambicchia, Enzo Maria Merlino, Berardo Ruffini, Matteo Talluri · 2025
Read Barycentric stability of nonlocal perimeters: the convex case by Chiara Gambicchia, Enzo Maria Merlino, Berardo Ruffini, Matteo Talluri (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-2506-03776.txt.
Introduction
Chiara Gambicchia and co-authors prove a sharp stability law for nonlocal perimeters of convex shapes in a 2025 paper. The classic isoperimetric law says round shapes enclose the most area for their boundary. The nonlocal twist replaces boundary length with a long-range interaction energy. For convex sets, the team tie the energy surplus to how far the shapes center of mass sits from its ideal spot, with best-possible constants. The result mirrors a known 1993 theorem, now in fractional dress.
Catalog record
| CREATOR | Chiara Gambicchia, Enzo Maria Merlino, Berardo Ruffini, Matteo Talluri |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 25 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-03776
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Barycentric stability of nonlocal perimeters: the convex case Chiara Gambicchia, Enzo Maria Merlino, Berardo Ruffini, Matteo Talluri Abstract . In this work, we establish a sharp form of a nonlocal quantitative isoperimetric inequality involving the barycentric asymmetry for convex sets. This result can be seen as the nonlocal analogue of the one obtained by Fuglede in [ 16 ] . 2020 Mathematics Subject Classification Primary 52A38, 49K40, 52A05; Secondary 28A75, 49Q10. 1. Introduction Quantitative isoperimetric inequalities have recently attracted considerable interest. The fundamental question of this research line is simple to state. Since it is well known that, among all sets with a given volume, the ball uniquely minimizes the perimeter, one may ask whether a set that nearly minimizes the perimeter must be itself close, in some precise sense, to a ball. Therefore, the goal is to establish a quantitative relation linking the perimeter excess of a set to its geometric proximity to a ball. To formalize this, we recall the notion of the isoperimetric deficit of a set E\subseteq\mathbb{R}^{n} , defined as \delta(E):=\frac{P(E)-P(B(m))}{P(B(m))}\,, where P(\cdot) is the perimeter in the sense of De Giorgi and B(m) denotes the ball centered at the origin with volume m=|E| . We point out that it is, by definition, scale invariant. Next, we need an index that measures how far a set is from being a ball, which leads to the definition of asymmetry . In literature, different notions of asymmetry have been introduced, all of which are scale invariant by definition. A possible choice is the Hausdorff asymmetry , defined by \lambda_{H}(E):=\inf\left\{\frac{d\big(E,(x+B(m))\big)}{|E|}:\,x\in\mathbb{R}^{n}\right\}, where d denotes the Hausdorff distance; see, e.g., [ 17 , Subsection 3.2] . Indeed, after some first contributions in the planar case [ 1 , 3 ] , in higher dimensions Fuglede [ 15 ] proved that for any convex set E , up to explicit multiplicative constants depending on the dimension, one can estimate the Hausdorff asymmetry by a suitable power of the deficit \delta(E) , with the correct order of magnitude as \delta\to 0 . However, it is not difficult to recognize that the Hausdorff asymmetry is too strong a notion when dealing with general sets of finite perimeter; see, for instance, [ 17 , Section 4] for an explicit counterexample. As first recognized by Hall in [ 20 ] , the Hausdorff asymmetry can be replaced by the Fraenkel asymmetry index : (1.1) \lambda(E):=\min\left\{\frac{\big|E\triangle(x+B(m))\big|}{|E|}:\,x\in\mathbb{R}^{n}\right\}, where we denote by " \triangle " the symmetric difference, that is, for any pair of sets A and B in \mathbb{R}^{n} , A\triangle B:=(A\setminus B)\cup(B\setminus A) . With this notion of asymmetry, as first proved in [ 18 ] , the sharp quantitative isoperimetric inequality states that, for any set E of finite perimeter, (1.2) \lambda(E)\leq C_{F}(n)\sqrt{\delta(E)}, where C_{F}(n) is a constant depending only on the dimension n . Inequality ( 1.2 ) has then been reproved with different techniques; see, for instance, [ 13 , 7 , 8 ] . Another notion of asymmetry, which is the one we focus on in this article, is the so-called barycentric asymmetry , defined by (1.3) \lambda_{0}(E):=\frac{\big|E\triangle(\operatorname{bar}(E)+B(m))\big|}{|E|}, where \operatorname{bar}(E) denotes the barycenter of E . So, while with the Fraenkel asymmetry the optimal ball is chosen to minimize the volume of the symmetric difference with the set, in the case of the barycentric asymmetry the ball is simply the one centered at the barycenter of the set. This is a strong and somehow arbitrary choice; however, it is reasonable to expect that in most cases if a set E is very close to a ball, then the center of this ball cannot be too far from the barycenter of E . Working with the barycentric asymmetry thus becomes very convenient, since it avoids the need for an optimization procedure. It is clear that, for instance in numerical approximations, it is computationally much more efficient to compute the barycenter and then evaluate the volume of the symmetric difference, rather than performing a minimization process. The corresponding sharp quantitative inequality, proved by Fuglede in [ 16 ] , reads as (1.4) \lambda_{0}(E)\leq C_{B}(N)\sqrt{\delta(E)}, whenever E is a convex set of finite perimeter. Observe that, as in ( 1.2 ) and unlike the case of Hausdorff asymmetry, the sharp exponent in the deficit is again \frac{1}{2} . As with the Hausdorff asymmetry, also with the barycentric one the inequality ( 1.4 ) is not valid for general sets and the same counterexample can be used to prove it, see [ 19 , Section 1] . However, recently, the estimate ( 1.4 ) has been extended under weaker geometric assumptions. In [ 2 ] , the authors showed that there exists a universal constant C_{BCH} such that, for every connected set E\subset\mathbb{R}^{2} , one has \lambda_{0}(E)\leq C_{BCH}\sqrt{\delta(E)}. Moreover, in [ 19 ] it was proved that, for every n\geq 2 and every D>0 , there exists a constant C(n,D) such that, for any set E\subseteq\mathbb{R}^{n} with \operatorname{diam}(E)\leq D|E|^{1/n} , it holds \lambda_{0}(E)\leq C(n,D)\sqrt{\delta(E)}. Motivated by these results, in this paper we focus on convex sets and aim to extend the barycentric quantitative isoperimetric inequality ( 1.4 ) to the fractional setting. In [ 6 ] , a notion of nonlocal perimeter was introduced, and the study of the corresponding minimizers was initiated. The simple yet profound idea underlying this new definition is to consider pointwise interactions between a set and its complementary, modulated by a kernel. The prototypical example involves singular kernels with polynomial decay. Concretely, given s\in(0,1) , the s –perimeter of a measurable set E\subseteq\mathbb{R}^{n} is defined as (1.5) P_{s}(E):=\int_{E}\int_{E^{c}}\frac{dx\,dy}{|x-y|^{n+s}}. The study of fractional perimeters is motivated by several applications. They naturally arise as nonlocal generalizations of the classical perimeter, interpolating between the Lebesgue measure and the De Giorgi’s perimeter functional, see, e.g., [ 23 ] . Moreover, fractional perimeters appear in models with long-range interactions, such as phase transitions, dislocation dynamics, and nonlocal diffusion processes, see [ 5 ] . The isoperimetric property of balls for the nonlocal perimeter was established in [ 14 ] : for any measurable set E\subset\mathbb{R}^{n} with |E|={|B(m)|} , one has (1.6) P_{s}({B(m)})\leq P_{s}(E), with equality if and only if E is a ball. The sharp quantitative version of the isoperimetric inequality ( 1.6 ) was later proved in [ 12 ] . For every n\geq 2 and s_{0}\in(0,1) , there exists a positive constant C(n,s_{0}) such that (1.7) \lambda(E)\leq C(n,s_{0})\sqrt{\delta_{s}(E)}, whenever s\in[s_{0},1] and 0<|E|<\infty , where (1.8) \delta_{s}(E):=\frac{P_{s}(E)-P_{s}({B(m)})}{P_{s}({B(m)})},\qquad\text{with }{|B(m)|}=|E|, is the s –isoperimetric deficit. The aim of this paper is to initiate the study of barycentric quantitative isoperimetric inequalities in the fractional setting, focusing on convex sets. More precisely, we establish a lower bound for the s –isoperimetric deficit in terms of the barycentric asymmetry, valid for any convex body in \mathbb{R}^{n} . Our main results is the following. Theorem 1.1 . For any s\in(0,1) and any convex body E\subseteq\mathbb{R}^{n} with finite measure and nonempty interior, there exists a constant C , depending only on n and s , such that (1.9) \lambda_{0}(E)\leq C\,\sqrt{\delta_{s}(E)}. Although in this note we focus on convex sets, the proof works as well for nearly spherical sets, see Remark 4.1 . This is a key point for extending the result to a