Generalized affine Hardy-Littlewood-Sobolev inequalities Journal: Journal of Functional Analysis
Youjiang Lin, Jinghong Zhou, Jiaming Lan · 2025
Read Generalized affine Hardy-Littlewood-Sobolev inequalities Journal: Journal of Functional Analysis by Youjiang Lin, Jinghong Zhou, Jiaming Lan (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-2508-01176.txt.
Introduction
A sharper shape-aware bound for a classic function inequality is proven by Youjiang Lin, Jinghong Zhou, and Jiaming Lan in a 2025 paper. The Hardy-Littlewood-Sobolev inequality caps how two functions can interact through a decaying kernel. Their affine version covers two different functions at once and beats the classic bound. Reverse forms of the new bounds hold for log-concave functions. Affine upgrades matter since they stay unchanged under volume-preserving shape stretches, which sharpens use in shape analysis.
Catalog record
| CREATOR | Youjiang Lin, Jinghong Zhou, Jiaming Lan |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 68 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-01176
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Generalized affine Hardy-Littlewood-Sobolev inequalities Journal: Journal of Functional Analysis Youjiang Lin, Jinghong Zhou, Jiaming Lan Abstract We establishe an affine Hardy-Littlewood-Sobolev inequality concerning two different functions which is stronger than the classical Hardy-Littlewood-Sobolev inequality. Furthermore, we also prove reverse inequalities for the new inequalities for log-concave functions. 2000 AMS subject classification: 26D15 (26B25, 26D10, 46E35, 52A40) Hardy–Littlewood–Sobolev inequality [ 21 ] states that for p,r>1 , 0<\alpha<n , \frac{1}{p}+\frac{1}{r}-\frac{\alpha}{n}=1 , the following inequality holds for functions f\in{L}^{p}(\mathbb{R}^{n}) and h\in{L}^{r}(\mathbb{R}^{n}) : \displaystyle C(n,\alpha,p){||f||}_{p}{||h||}_{r}\geq\left|\int_{\mathbb{R}^{n}}\int_{\mathbb{R}^{n}}\frac{f(x)h(y)}{{|x-y|}^{n-\alpha}}dxdy\right|, (1) where C(n,\alpha,p) is a sharp constant independent of f and h and satisfies \displaystyle C(n,\alpha,p)\leq\frac{n}{\alpha}\omega_{n}^{1-\frac{\alpha}{n}}\frac{1}{pr}\left({\left(\frac{1-\frac{\alpha}{n}}{1-1/p}\right)}^{1-\frac{\alpha}{n}}+{\left(\frac{1-\frac{\alpha}{n}}{1-1/r}\right)}^{1-\frac{\alpha}{n}}\right), (2) where \omega_{n} represents the volume of the unit ball in \mathbb{R}^{n} . In the special case where p=r=2n/(n+\alpha) , the constant C(n,\alpha,p) simplifies to: \displaystyle C(n,\alpha,p)=C(n,\alpha)={\pi}^{(n-\alpha)/2}\frac{\Gamma(\alpha/2)}{\Gamma((n+\alpha)/2)}{\left\{\frac{\Gamma(n/2)}{\Gamma(n)}\right\}}^{-\alpha/n}, where \Gamma is the gamma function. Equality in ( 1 ) holds if and only if h=a_{0}f and f(x)=a(b^{2}+|x-x_{0}|^{2})^{-\frac{n+\alpha}{2}} for some a_{0} , a\in\mathbb{R} , 0\neq b\in\mathbb{R} and x_{0}\in\mathbb{R}^{n} . The Hardy–Littlewood-Sobolev inequality was initially introduced by G. Hardy and J. Littlewood in \mathbb{R}^{1} [ 17 , 18 , 19 ] , and was later extended by S. Sobolev [ 35 ] to \mathbb{R}^{n} . However, neither of these inequalities is in its optimal form. Specifically, neither the best possible constant nor the extremal functions that would make the inequality (1.1) hold with this constant were identified. In a particular condition, E. Lieb [ 20 ] provided the optimal version of (1.1) by proving the inequality with the best possible constant and identifying the functions that achieve equality with this constant. For the general case, neither the optimal constant nor the optimizers are known. For more detailed information on the Hardy–Littlewood–Sobolev inequality, we can refer to the monograph by E. Lieb and M. Loss [ 21 ] . For geometric questions, affine inequalities are unchanged under translations and volume-preserving linear transformations. The most prominent example is the Petty projection inequality for convex bodies (i.e., compact convex sets) in \mathbb{R}^{n} , which not only strengthens the Euclidean isoperimetric inequality but actually implies it (see, for example, [ 3 , 6 , 8 , 24 , 34 ] , and see [ 1 , 5 , 12 , 13 , 15 , 16 , 28 , 29 , 22 , 23 , 30 , 32 , 33 , 36 , 37 , 39 ] for related results in the {L}_{p} Brunn–Minkowski theory and for more general sets). J. Haddad and M. Ludwig [ 14 , Theorem 1] established sharp affine Hardy-Littlewood-Sobolev inequalities for 0<\alpha<n and non-negative f\in{L}^{2n/(n+\alpha)}(\mathbb{R}^{n}) : \displaystyle\gamma_{n,\alpha}\left\|f\right\|_{\frac{2n}{n+\alpha}}^{2} \displaystyle\geq \displaystyle n\omega_{n}^{\frac{n-\alpha}{n}}\left(\frac{1}{n}\int_{\mathbb{S}^{n-1}}\left(\int_{0}^{\infty}t^{\alpha-1}\int_{\mathbb{R}^{n}}f\left(x\right)f\left(x+t\xi\right)dxdt\right)^{\frac{n}{\alpha}}d\xi\right)^{\frac{\alpha}{n}} (3) \displaystyle\geq \displaystyle\int_{\mathbb{R}^{n}}\int_{\mathbb{R}^{n}}\frac{f\left(x\right)f\left(y\right)}{\left|x-y\right|^{n-\alpha}}dxdy. Equality holds in the first inequality if and only if the function f has the form f\left(x\right)=a\left(1+\left|\phi\left(x-x_{0}\right)^{2}\right|\right)^{-\frac{n+\alpha}{2}} for x\in\mathbb{R}^{n} , where a\geq 0 , \phi\in GL\left(n\right) and x_{0}\in\mathbb{R}^{n} . Here GL(n) denotes general linear transformations. For the second inequality, equality is attained precisely if f is radially symmetric. Here, the integral over the unit sphere \mathbb{S}^{n-1} is taken with respect to the (n-1) -dimensional Hausdorff measure. The sharp affine Hardy-Littlewood-Sobolev inequalities ( 3 ) are significantly stronger than the classical Hardy-Littlewood-Sobolev inequality ( 1 ) for f=h . The primary goal of this paper focuses on proving affine Hardy-Littlewood-Sobolev inequalities concerning two different functions. Theorem 1 . For n\geq 1 , 1<p,r<\infty , and 0<\alpha<n satisfying \displaystyle\frac{1}{p}+\frac{1}{r}-\frac{\alpha}{n}=1. Then, there is a constant C\left(n,\alpha,p\right)>0 such that for all non-negative functions f\in L^{p}\left(\mathbb{R}^{n}\right) and h\in L^{r}\left(\mathbb{R}^{n}\right) , \displaystyle C\left(n,\alpha,p\right)\left\|f\right\|_{p}\left\|h\right\|_{r} \displaystyle\geq \displaystyle n\omega^{\frac{n-\alpha}{n}}_{n}\left(\frac{1}{n}\int_{\mathbb{S}^{n-1}}\left(\int_{0}^{\infty}t^{\alpha-1}\int_{\mathbb{R}^{n}}h\left(x\right)f\left(x+t\xi\right)dxdt\right)^{\frac{n}{\alpha}}d\xi\right)^{\frac{\alpha}{n}} (4) \displaystyle\geq \displaystyle\int_{\mathbb{R}^{n}}\int_{\mathbb{R}^{n}}\frac{f\left(x\right)h\left(y\right)}{\left|x-y\right|^{n-\alpha}}dxdy. If p=r=\frac{2n}{n+\alpha} , there is equality in the first inequality if and only if h coincides with a translation of a_{0}f where a_{0}\in\mathbb{R} and \displaystyle f\left(x\right)=a\left(b^{2}+\left|\phi\left(x-x_{0}\right)^{2}\right|\right)^{-\frac{n+\alpha}{2}} for some a\in\mathbb{R} , 0\neq b\in\mathbb{R} , \phi\in GL\left(n\right) and x_{0}\in\mathbb{R}^{n} . There is equality in the second inequality if f , h are radially symmetric. The generalized affine Hardy–Littlewood–Sobolev inequality ( 4 ) is stronger than Hardy–Littlewood–Sobolev inequality ( 1 ). When p=r=\frac{2n}{n+\alpha} and f=h , we can get the affine Hardy-Littlewood-Sobolev inequalities ( 3 ). To prove Theorem 1 , for given 0<\alpha<n , we define the star-shaped set S_{\alpha}\left(f,h\right) associated with the functions f and h by its radial function, \displaystyle\rho^{\alpha}_{S_{\alpha}\left(f,h\right)}\left(\xi\right)=\int_{0}^{\infty}t^{\alpha-1}\int_{\mathbb{R}^{n}}h\left(x\right)f\left(x+t\xi\right)dxdt,\hskip 10.00002pt\xi\in\mathbb{R}^{n}\setminus\{0\}. (5) The first inequality presented in Theorem 1 now can be expressed in the form \displaystyle C\left(n,\alpha,p\right)\left\|f\right\|_{p}\left\|h\right\|_{r}\geq n\omega^{\frac{n-\alpha}{n}}_{n}\left|S_{\alpha}\left(f,h\right)\right|^{\frac{\alpha}{n}}, (6) where \left|\cdot\right| denotes n -dimensional Lebesggue measure. Since both sides of ( 6 ) remain invariant under same translations of f and h , and that \displaystyle S_{\alpha}\left(f\circ\phi^{-1},h\circ\phi^{-1}\right)=\phi S_{\alpha}\left(f,h\right) holds for volume-preserving linear transformations \phi:\mathbb{R}^{n}\to\mathbb{R}^{n} , it follows that the inequality ( 6 ) is affine. J. Dou and M. Zhu [ 7 ] and W. Beckner [ 2 ] obtained sharp Hardy-Littlewood-Sobolev inequalities for \alpha>n as follows: \displaystyle C\left(n,\alpha,p\right)\left\|f\right\|_{p}\left\|h\right\|_{r}\leq\int_{\mathbb{R}^{n}}\int_{\mathbb{R}^{n}}\frac{f\left(x\right)h\left(y\right)}{\left|x-y\right|^{n-\alpha}}dxdy (7) for non-negative functions f\in L^{p}\left(\mathbb{R}^{n}\right) and h\in L^{r}\left(\mathbb{R}^{n}\right) satisfying \frac{1}{p}+\frac{1}{r}-\frac{\alpha}{n}=1 , where C\left(n,\alpha,p\right) is defined in ( 2 ). When p=r=\frac{2n}{n+\alpha} , equality in the inequality holds if and only if h=a_{0}f and \displaystyle f\left(x\right)=a\left(b^{2}+\left|\left(x-x_{0}\right)^{2}\right|\right)^{-\frac{n+\alpha}{2}} for some a_{0} , a\in\mathbb{R} , 0\neq b\in\mathbb{R} , and x_{0}\in\mathbb{R}^{n} . In addtion, J. Haddad and M. Ludwig [ 14 , Theorem 14] etablished corresponding sharp affine Hardy-Littlewood-Sobolev inequalities for \alpha>n and non-negative f\in{L}^{2n/(n+\alpha)}(\mathbb{R}^{n}) : \displaystyle\gamma_{n,\alpha}\left\|f\right\|_{\frac{2n}{n+\alpha}}^{2}\leq n\omega_{n}^{\frac{n-\alpha}{n}}\left|S_{\alpha}f\right|^{\frac{\alpha}{n}}\leq\int_{\mathbb{R}^{n}}\int_{\mathbb{R}^{n}}\frac{f\left(x\right)f\left(y\right)}{\left|x-y\right|^{n-\alpha}}dxdy. (8) Equality holds in the first inequality if and only if the function f has the form f\left(x\right)=a\left(1+\left|\phi\left(x-x_{0}\right)^{2}\right|\right)^{-\frac{n-\alpha}{2}} for x\in\mathbb{R}^{n} , where a\geq 0 , \phi\in GL\left(n\right) and x_{0}\in\mathbb{R}^{n} . For the second inequality, equality is attained precisely if f is radially symmetric. In Section 4 , for \alpha>n , we will establish affine HLS inequalities concerning two different functions. Theorem 2 . For n\geq 1 , 0<p,r<1 , and \alpha>n satisfying \displaystyle\frac{1}{p}+\frac{1}{r}-\frac{\alpha}{n}=1, there is a constant C\left(n,\alpha,p\right)>0 such that for all non-negative f\in L^{p}\left(\mathbb{R}^{n}\right) and h\in L^{r}\left(\mathbb{R}^{n}\right) , \displaystyle C\left(n,\alpha,p\right)\left\|f\right\|_{p}\left\|h\right\|_{r}\leq n\omega^{\frac{n-\alpha}{n}}_{n}\left|S_{\alpha}\left(f,h\right)\right|^{\frac{\alpha}{n}}\leq\int_{\mathbb{R}^{n}}\int_{\mathbb{R}^{n}}\frac{f\left(x\right)h\left(y\right)}{\left|x-y\right|^{n-\alpha}}dxdy. (9) If p=r=\frac{2n}{n+\alpha} , there is equality in the first inequality if and only if h coincides with a translation of a_{0}f where a_{0}\in\mathbb{R} and \displaystyle f\left(x\right)=a\left(b^{2}+\left|\phi\left(x-x_{0}\right)^{2}\right|\right)^{-\frac{n+\alpha}{2}} for some a\in\mathbb{R},0\neq b\in\mathbb{R} , \phi\in GL\left(n\right) and x_{0}\in\mathbb{R}^{n} . There is equality in the second inequality if f , h are radially symmetric. The generalized affine Hardy–Littlewood–Sobolev inequality ( 9 ) is stronger than sharp Hardy–Littlewood–Sobolev inequality ( 7 ). When p=r=\frac{2n}{n+\alpha} and f=h , we can get the sharp affine Hardy-Littlewood-Sobolev inequalities ( 8 ). In Section 5 , for measurable sets F\subseteq E\subseteq\mathbb{R}^{n} , we will consider S_{\alpha}\left(1_{E},1_{F}\right) , where 1_{E} and 1_{F} denote the indicator functions of E and F , respectively. For n -dimensional convex bodies F\subseteq E\subseteq\mathbb{R}^{n} , we demonstrate that S_{\alpha}\left(1_{E},1_{F}\right) is proportional to the generalized radial \alpha -mean body of E and F , a significant concept introduced by R. Gardner and G. Zhang [ 10 , p.524] . In [ 16 ] , sharp isoperinetric inequalities were recently proved for radial \alpha -mean bodies, which is special case of generalized radial \alpha -mean body for E=F . As shown by R. Gardner and G. Zhang [ 10 ] , sharp reverse inequalities hold. They extend Zhang’s reverse Petty projection inequality [ 38 ] , with equality being achieved exactly for n -dimensional simplices. Moreover, we introduce the conceptions of radial mean bodies in term of functions. For 0<\alpha<n , J. Haddad and M. Ludwig [ 14 , Theorem 2] etablished the sharp reverse inequalities for sharp affine HLS inequalities ( 3 ) as follows: \displaystyle\frac{\Gamma\left(n+1\right)^{\frac{\alpha}{n}}}{\Gamma\left(\alpha\right)}\left|S_{\alpha}f\right|^{\frac{\alpha}{n}}\geq{\left\|f\right\|}_{2}^{2-\frac{2\alpha}{n}}{\left\|f\right\|}_{1}^{\frac{2\alpha}{n}}\geq{\left\|f\right\|}_{\frac{2n}{n+\alpha}}^{2} (10) for