Sharp quantitative Talenti’s inequality in particular cases

P. Acampora, J. Lamboley · 2025

Read Sharp quantitative Talenti’s inequality in particular cases by P. Acampora, J. Lamboley (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-2503-07337.txt.

Introduction

Paolo Acampora and Jimmy Lamboley sharpen in 2025 a classic inequality about how smoothing a function affects the size of its solution. The classic result compares a heat like equation on a domain with its round, symmetric twin. Stability asks how near the twin a near equal case must be, and with what power.

The authors take one clean case: functions on a unit ball that mark out a subset, one inside and zero outside. There they prove stability with the sharp power of two. Their proof pins down the best possible rate for this case. The work adds a firm brick to a long studied wall.

Catalog record

CREATORP. Acampora, J. Lamboley
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE120 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2503-07337

Browse the full library · About the Tower of Babel

Full text

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

Sharp quantitative Talenti’s inequality in particular cases P. Acampora, J. Lamboley Abstract In this paper, we focus on the famous Talenti’s symmetrization inequality, more precisely its L p L^{p} corollary asserting that the L p L^{p} -norm of the solution to − Δ ​ v = f ♯ -\Delta v=f^{\sharp} is higher than the L p L^{p} -norm of the solution to − Δ ​ u = f -\Delta u=f (we are considering Dirichlet boundary conditions, and f ♯ f^{\sharp} denotes the Schwarz symmetrization of f : Ω → ℝ + f:\Omega\to\mathbb{R}_{+} ). We focus on the particular case where functions f f are defined on the unit ball, and are characteristic functions of a subset of this unit ball. We show in this case that stability occurs for the L p L^{p} -Talenti inequality with the sharp exponent 2. 1 Introduction In this paper, we will investigate some stability versions of Talenti’s inequality in particular cases. Given \Omega\subset\mathbb{R}^{n} a bounded open set and f\in L^{2}(\Omega) , we will denote by u_{f} the unique weak solution to \left\{\begin{array}[]{ccccc}-\Delta u_{f}&=&f&\textrm{ in }&\Omega,\\ u_{f}&=&0&\textrm{ on }&\partial\Omega.\end{array}\right. (1.1) The aim of Talenti’s inequality is to compare u_{f} and u_{f^{\sharp}} where f^{\sharp}:\Omega^{\sharp}\to\mathbb{R} is the Schwarz symmetrization of f , defined on \Omega^{\sharp} the centered ball of same volume as \Omega (see Definition 2.1 ): more precisely Talenti’s inequality (Theorem 2.4 ) states that if f\geq 0 , then \forall x\in\Omega^{\sharp},\;\;u_{f}^{\sharp}(x)\leq v(x), where v=u_{f^{\sharp}} solves \left\{\begin{array}[]{ccccc}-\Delta v&=&f^{\sharp}&\textrm{ in }&\Omega^{\sharp}\\ v&=&0&\textrm{ on }&\partial\Omega^{\sharp}.\end{array}\right. (1.2) and u_{f}^{\sharp} is the Schwarz symmetrization of u_{f} . This ponctual comparison implies in particular the following: for every p\in[1,+\infty] , \|u_{f}\|_{L^{p}(\Omega)}\leq\|v\|_{L^{p}(\Omega^{\sharp})}. (1.3) Moreover, as shown in [ 2 ] (see Theorem 2.5 ), equality is realized in ( 1.3 ) for some p\in[1,+\infty] only if \Omega is a ball and f=f^{\sharp} up to translations, see also [ 16 ] . In this paper we are interested in quantitative versions of ( 1.3 ). Notice that there are two parameters that are symmetrized in these inequalities: \Omega and f . Therefore the stability inequalities we may be seeking for should take into account both the distance from f to f^{\sharp} and from \Omega to \Omega^{\sharp} : to evaluate the asymmetry of \Omega for example, we denote \alpha(\Omega)=\min_{x_{0}\in\mathbb{R}^{n}}\left\{\frac{|\Omega\Delta B_{r}(x_{0})|}{|B_{r}|},\;|B_{r}|=|\Omega|\right\} the Fraenkel asymmetry of \Omega ( B_{r}(x) denotes the ball of radius r>0 and centered at x\in\mathbb{R}^{n} ; we also denote B_{r}=B_{r}(0) the centered ball of radius r ). As far as we know, there are only two partial results in this direction: 1. in [ 9 ] , the author focuses on the case f\equiv 1 and shows that for every p\in[1,+\infty] , there exists c=c(n,p) such that for every \Omega bounded open set in \mathbb{R}^{n} , \begin{split}&\|v\|^{p}_{L^{p}(\Omega^{\sharp})}-\|u_{1}\|^{p}_{L^{p}(\Omega)}\geq c\alpha(\Omega)^{2+p}\qquad\textrm{ if }p\in[1,+\infty),\\ &\|v\|_{L^{\infty}(\Omega^{\sharp})}-\|u_{1}\|_{L^{\infty}(\Omega)}\geq c\alpha(\Omega)^{3}.\end{split} 2. in [ 3 ] the authors focus on the case p=\infty : it is shown that for any \Omega bounded open set of \mathbb{R}^{n} and any nonnegative f\in L^{2}(\Omega) , there exists c=c(n,|\Omega|,f^{\sharp}) and \theta=\theta(n) such that \|v-u^{\sharp}\|_{L^{\infty}(\Omega^{\sharp})}\geq c\left(\alpha(\Omega)^{3}+\inf_{x_{0}\in\mathbb{R}^{n}}\Big\|f-f^{\sharp}(\cdot+x_{0})\Big\|_{L^{1}(\mathbb{R}^{n})}^{\theta}\right). Note that the proof of Talenti’s inequality shows that v-u^{\sharp} is radially decreasing, so that \lVert v-u^{\sharp}\rVert_{L^{\infty}(\Omega^{\sharp})}=\lVert v\rVert_{L^{\infty}(\Omega^{\sharp})}-\lVert u\rVert_{L^{\infty}(\Omega)} . In the same paper, the authors also obtain a partial result in the case p=2 , namely: \|v\|_{L^{2}(\Omega^{\sharp})}^{2}-\|u\|_{L^{2}(\Omega)}^{2}\geq c\alpha(\Omega)^{4}. In both of these results, the obtained exponents are not expected to be sharp (also in the second result, \theta does not seem to be explicit). The goal of this paper is to provide a first quantitative result for ( 1.3 ) with a sharp exponent. In order to start this investigation with only one parameter (similarly to [ 9 ] where the author assumed f\equiv 1 ), we will assume that \Omega=B_{1} is a centered ball of radius 1, so that \Omega=\Omega^{\sharp} and the only parameter is f . Moreover, we will assume f=\chi_{E} to be the characteristic function of a set E\subset B_{1} : in this case we denote u_{E} for u_{\chi_{E}} , and f^{\sharp}=\chi_{B_{*}} where B_{*} is the centered ball of same volume as E . This will allow us to use a geometric approach and the framework of shape derivatives. In this setting, we obtain the following stability result: Theorem 1.1 . Let n\geq 1 , p\in[1,+\infty] , and m\in(0,|B_{1}|) . Then there exists c=c(p,m,n)>0 such that for every measurable set E\subset B_{1} with \lvert E\rvert=m we have \lVert u_{B_{*}}\rVert_{L^{p}(B_{1})}-\lVert u_{E}\rVert_{L^{p}(B_{1})}\geq{c}\lvert E\Delta B_{*}\rvert^{2}, (1.4) where B_{*} is the centered ball of volume m . Moreover, we show in Propositions 2.2 , 2.4 and 2.4 that the exponent 2 obtained in Theorem 1.1 is sharp. In a forthcoming work [ 1 ] we will generalize the strategy introduced in the present paper to investigate the more general case where f is not assumed to be a characteristic function. The proof of Theorem 1.1 will be divided in three cases depending on the value of p . In the special case p=1 , the proof is quite straightforward and relies only on Talenti’s inequality: we provide the proof in Section 2.2 , which is based on the study of the following shape optimization problem: \max\Set{\lVert u_{E}\rVert_{1}}{\begin{gathered}|E|=m,\\ |E\Delta B_{*}|=\delta\end{gathered}} for m\in(0,|B_{1}|) and \delta>0 , which can be seen as looking for the worst set of given asymmetry, with regard of the Talenti deficit \|u_{B_{*}}\|_{1}-\|u_{E}\|_{1} . We identify explicitely the solution of this problem as a target-shaped set (it is the union of a ball and an annulus, see Section 2.2 ). In the case p>1 , we were not able to solve explicitly \max\{\|u_{E}\|_{p},|E|=m,|E\Delta B_{*}|=\delta\} , we therefore follow a strategy developed in [ 18 , 19 , 5 , 6 ] to prove ( 1.4 ). In fact the case 1<p<\infty falls into the following framework: given a function j:\mathbb{R}_{+}\to\mathbb{R} , we study the stability for the following shape optimization problem: \max\Set{\mathcal{J}(E)}{\lvert E\rvert=m},\qquad\qquad\mathcal{J}(E):=\int_{B_{1}}j(u_{E}(x))dx (1.5) which is solved by the centered ball B_{*} if j is non-decreasing. Classically, under convexity assumption on j , such problem can be relaxed in the class \mathcal{M}_{m}:=\Set{V\in L^{\infty}(B_{1})}{\begin{gathered}0\leq V\leq 1,\\ \int_{B_{1}}V=m\end{gathered}} where m\in(0,|B_{1}|) , and we define \mathcal{J}(V):=\int_{B_{1}}j(u_{V}) for V\in L^{\infty}(B_{1}) non-negative. More precisely we prove: Theorem 1.2 . Let n\geq 1 , m\in(0,\lvert B_{1}\rvert) , and j\in C^{1}(\mathbb{R}_{+})\cap C^{2}((0,+\infty)) be such that j^{\prime}(0)\geq 0 and j^{\prime\prime}(s)>0 for every s>0 , and such that \limsup_{s\to 0^{+}}s^{\alpha}j^{\prime\prime}(s)<+\infty (1.6) for some \alpha<1 . Then there exists a positive constant c=c(j,m,n) such that for every V\in\mathcal{M}_{m} we have \mathcal{J}(B_{*})-\mathcal{J}(V)\geq c\lVert V-\chi_{B_{*}}\rVert_{1}^{2}. where B_{*} is the centered ball of volume m . Remark 1.3 . Note that assumption ( 1.6 ) implies that j^{\prime} is locally \beta:=(1-\alpha) -Hölder continuous. This follows easily from the equality j^{\prime}(y)-j^{\prime}(x)=\int_{x}^{y}j^{\prime\prime}(t)dt . Proof of 1.1 from 1.2 when p\in(1,\infty) : Let p\in(1,\infty) : we consider j:s\in[0,\infty)\mapsto s^{p} , which is C^{2} in (0,\infty) and C^{1} up to 0, with j^{\prime}(0)=0 and j^{\prime\prime}>0 on (0,\infty) , and satisfies ( 1.6 ). Therefore, 1.2 applies, and there exists c=c(p,m,n) such that \lVert u_{B_{*}}\rVert_{p}^{p}\left(1-\left(\frac{\lVert u_{E}\rVert_{p}}{\lVert u_{B_{*}}\rVert_{p}}\right)^{p}\right)\geq c\lvert E\Delta B_{*}\rvert^{2}. Then the result follows by using the inequality 1-x^{p}\leq p(1-x)\qquad\qquad\forall x\geq 0. ∎ Finally, we give a seperate proof of Theorem 1.1 when p=\infty , see below for more details. Outline of the paper In the following section, we recall basic definitions on Schwarz symmetrization, and then

Read the complete text