The Gaussian-Minkowski problem for C C -pseudo-cones

Xudong Wang, Tingting Xiang · 2024

Read The Gaussian-Minkowski problem for C C -pseudo-cones by Xudong Wang, Tingting Xiang (2024) 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-2412-20908.txt.

Description

Abstract The Gaussian surface area measures for C C -pseudo-cones are studied in this paper. Using the variational arguments and the approximation methods of Schneider, we obtain the existence of solutions to the Gaussian-Minkowski problem for C C -pseudo-cones with small co-volume. Mathematics Subject Classification 2020. 52A20, 52A30, 52A40. Keywords. Gaussian-Minkowski problem, C C -pseudo-cone.

Catalog record

CREATORXudong Wang, Tingting Xiang
YEAR2024
COLLECTIONPAPERS
FORMATTXT
FILE SIZE30 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2412-20908

Browse the full library · About the Tower of Babel

Full text

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

The Gaussian-Minkowski problem for C C -pseudo-cones Xudong Wang, Tingting Xiang Abstract The Gaussian surface area measures for C C -pseudo-cones are studied in this paper. Using the variational arguments and the approximation methods of Schneider, we obtain the existence of solutions to the Gaussian-Minkowski problem for C C -pseudo-cones with small co-volume. Mathematics Subject Classification 2020. 52A20, 52A30, 52A40. Keywords. Gaussian-Minkowski problem, C C -pseudo-cone. This paper was supported by the Excellent Graduate Training Program of SNNU (No. LHRCCX23142). 1 Introduction The classical Minkowski problem asks for some convex body in the n -dimensional Euclidean space \mathbb{R}^{n} such that its surface area measure is a given finite Borel measure \mu on the unit sphere \mathbb{S}^{n-1} . This was solved by Minkowski [ 40 , 41 ] in the case of discrete data and continuous density. Aleksandrov [ 4 ] and Fenchel-Jessen [ 22 ] solved the general case of measure, respectively. Given a smooth function f on \mathbb{S}^{n-1} , then the smooth version of the Minkowski problem is a prescribed Gauss curvature problem, namely the following Monge-Ampère equation on \mathbb{S}^{n-1} : \det(\nabla^{2}h+hI)=f, where \nabla^{2} is the covariant Hessian operator on \mathbb{S}^{n-1} and I is the identity matrix. On this regularity of the Minkowski problem, one can refer to Lewy [ 32 , 33 ] , Nirenberg [ 42 ] , Pogorelov [ 43 ] , Cheng and Yau [ 16 ] , Caffarelli [ 13 , 14 ] and so on. Recent years, there appear many Minkowski type problems. Such as, L_{p} Minkowski problem [ 35 ] , Orlicz Minkowski problem [ 23 ] , dual Minkowski problem [ 25 ] , Gauss image problem [ 12 ] , Chord Minkowski problem [ 37 ] , Gaussian-Minkowski problem [ 27 ] , Minkowski problem for electrostatic capacity [ 30 ] , torsional rigidity [ 21 ] , harmonic measure [ 29 ] , and their various deformations and combinations. In this paper, we consider the Minkowski type problem for unbounded closed convex sets in the Gaussian probability space. Given a pointed closed convex cone C , Schneider [ 48 , 50 ] firstly studied the Minkowski problem for the surface area measures and the cone-volume measures of unbounded closed convex sets in C . Later, many Minkowski type problems for unbounded closed convex sets in Lebesgue measure space [ 62 , 34 , 15 , 3 , 56 , 63 , 55 , 51 ] were also solved. In [ 52 , 53 ] , Schneider studied the weighted Minkowski problem for C -pseudo-cones (see Section 2 for the definition), where the weighted functions have homogeneity and the weighted surface area measures for all C -pseudo-cones are finite. Following the same finiteness idea, the Minkowski type problems for C -pseudo-cones in the Gaussian probability space are attractive. The setting of this paper is the n -dimensional Euclidean sapce \mathbb{R}^{n} with the standard inner product \langle\cdot,\cdot\rangle and norm |\cdot| . We denote by \mathbb{S}^{n-1} the unit sphere in \mathbb{R}^{n} . We also denote by int and \partial the interior operator and the boundary operator, respectively. We fix a n -dimensional closed convex cone C in \mathbb{R}^{n} and suppose that C is pointed, i.e., C\cap(-C)=\{o\} . The polar cone of C is defined by C^{\circ}=\{x\in\mathbb{R}^{n}\,|\,\langle x,y\rangle\leqslant 0,\ \forall\ y\in C\}, and we denote \Omega_{C}=\text{{int}}\,C\cap\mathbb{S}^{n-1} , \Omega_{C^{\circ}}=\text{{int}}\,C^{\circ}\cap\mathbb{S}^{n-1} . Let E be a C -pseudo-cone. The Gaussian surface area measure of E is defined by S_{\gamma_{n}}(E,\omega)=\frac{1}{(2\pi)^{\frac{n}{2}}}\int_{\boldsymbol{\nu}^{*}_{E}(\omega)}e^{-\frac{|x|^{2}}{2}}\,d\mathcal{H}^{n-1}(x) for any Borel set \omega\subset\Omega_{C^{\circ}} . Here \boldsymbol{\nu}^{*}_{E} represents the inverse Gauss image of E and \mathcal{H}^{k} represents the k -dimensional Hausdorff measure in \mathbb{R}^{n} . In Lemma 3.1 , we verify the finiteness of the Gaussian surface area measures for all C -pseudo-cones. Moreover, the Gaussian co-volume of E is defined by \overline{\gamma}_{n}(E)=\gamma_{n}(C\setminus E)=\frac{1}{(2\pi)^{\frac{n}{2}}}\int_{C\setminus E}e^{-\frac{|x|^{2}}{2}}\,d\mathcal{H}^{n}(x), where \gamma_{n} is the Gaussian probability measure in \mathbb{R}^{n} . Naturally, the Gaussian-Minkowski problem for C -pseudo-cones can be stated as the following: Given a nonzero finite Borel measure \mu on \Omega_{C^{\circ}} , is there a C -pseudo-cone E such that the Gaussian surface area measure of E is \mu ? In the case of convex bodies, the Gaussian-Minkowski problem has been studied by Huang, Xi and Zhao [ 27 ] . As its unbounded analogue, the existence of some solutions to the Gaussian-Minkowski problem for C -pseudo-cones is confirmed. Here we adopt the methods of Schneider [ 48 , 52 ] and some technologies in [ 27 ] . By establishing some uniform estimates under the restriction of small co-volume, we obtain the following result: Theorem 1.1 . Let \mu be a nonzero Borel measure on \Omega_{C^{\circ}} , then there exists a C -pseudo-cone E with \overline{\gamma}_{n}(E)\leqslant\frac{1}{2}\gamma_{n}(C) such that S_{\gamma_{n}}(E,\cdot)=\mu if and only if \mu is finite. The organization of this paper is as follows. In Section 2, we introduce some tools regarding C -pseudo-cones. In Section 3, some properties of the Gaussian surface area measures for C -pseudo-cones are studied. Thanks to the uniform estimates in Section 4, the existence of some solutions on compact sets is established by the variational arguments. Finally, the proof of Theorem 1.1 is given in Section 5. 2 Preliminaries We denote by B^{n} the unit ball in \mathbb{R}^{n} , and denote by o the origin of \mathbb{R}^{n} . If we choose \mathfrak{u}\in\Omega_{C}\cap(-\Omega_{C^{\circ}}) , then \langle\mathfrak{u},x\rangle>0 for any x\in C\setminus\{o\} , see [ 51 ] . We set C(t)=\{x\in C\,|\,\langle x,\mathfrak{u}\rangle=t\} and C^{-}(t)=\{x\in C\,|\,\langle x,\mathfrak{u}\rangle\leqslant t\} for t>0 . Let o\notin E\subset C be a nonempty closed convex set, then E is called a pseudo-cone if \lambda x\in E\ \,\text{for any}\ \lambda\geqslant 1\ \text{and}\ x\in\,E. Moreover, E is called a C -pseudo-cone if {\rm rec}\,E=C , where {\rm rec}\,E is the recession cone of E defined by {\rm rec}\,E=\{x\in\mathbb{R}^{n}\,|\,E+x\subset E\}. We denote by \mathcal{PC}(C) the set of all C -pseudo-cones. Let E\in\mathcal{PC}(C) . If C\setminus E has finite volume, then E is called a C -close set and C\setminus E is called a C -coconvex set. In particular, a C -close set E is called a C -full set if C\setminus E is bounded. The support function of E is defined by h_{E}(v)=\sup\{\langle x,v\rangle\,|\,x\in E\},\,v\in\Omega_{C^{\circ}}. Note that -\infty<h_{E}(v)<0 for any v\in\Omega_{C^{\circ}} , one can define \overline{h}_{E}=-h_{E} , so that \overline{h}_{E} is a positive continuous function on \Omega_{C^{\circ}} . The radial function of E is defined by \rho_{E}(u)=\inf\{r>0\,|\,ru\in E\},\,u\in\Omega_{C}. Let \omega\subset\Omega_{C^{\circ}} be a nonempty compact set and K\in\mathcal{PC}(C) . K is called a C -determined set by \omega if K=C\cap\bigcap_{v\in\omega}H^{-}(v,h_{K}(v)). We denote by \mathcal{K}(C,\omega) the class of all C -determined sets by \omega . Note that a C -determined set by a compact set is also a C -full set. Moreover, one can define a C -determined set K^{(\omega)} of a C -pseudo-cone K by K^{(\omega)}=C\cap\bigcap_{v\in\omega}H^{-}(v,h_{K}(v)). Denote by \widetilde{\partial}E=\partial E\cap\text{{int}}\,C . We use \mathfrak{B}(X) to represent the class of all Borel sets in X . For E\in\mathcal{PC}(C) and \beta\in\mathfrak{B}(\widetilde{\partial}E) , the Gauss image of E is defined by \boldsymbol{\nu}_{E}(\beta)=\{v\in\Omega_{C^{\circ}}\,|\,\mbox{there exists}\ x\in\beta\ \mbox{such that}\ \langle x,v\rangle=h_{E}(v)\}, and for \omega\in\mathfrak{B}(\Omega_{C^{\circ}}) , the inverse Gauss image of E is defined by \boldsymbol{\nu}^{*}_{E}(\omega)=\{x\in\widetilde{\partial}E\,|\,\mbox{there exists}\ v\in\omega\ \mbox{such that}\ \langle x,v\rangle=h_{E}(v)\}. Let \{E_{i}\,|\,i\in\mathbb{N}_{0}\}\subset\mathcal{PC}(C) , we say that E_{i} converges to E_{0} as i\rightarrow+\infty in the sense of Schneider if there exists

Read the complete text