On the packing dimension of distance sets with respect to 𝑪 \bm{C^{1\!}} and polyhedral norms
Iqra Altaf, Ryan Bushling, Bobby Wilson · 2025
Read On the packing dimension of distance sets with respect to 𝑪 \bm{C^{1\!}} and polyhedral norms by Iqra Altaf, Ryan Bushling, Bobby Wilson (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-2504-11660.txt.
Introduction
Iqra Altaf and co-authors measure how rich the set of distances inside a fractal like set can be. Take a set in space and list all gaps between its pairs of points. How large is that gap set, in packing terms? The answer depends on the ruler used to measure the gaps.
For a wide family of rulers, smooth or faceted like a cut gem, the team proves a floor. The gap set's size is at least the set's own size divided by the dimension. A key tool stretches a projection result from earlier work. A build in the paper shows the floor is exact for many faceted rulers, so no stronger claim can hold in general.
Catalog record
| CREATOR | Iqra Altaf, Ryan Bushling, Bobby Wilson |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 36 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2504-11660
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
On the packing dimension of distance sets with respect to 𝑪 \bm{C^{1\!}} and polyhedral norms Thanks: B. W. is supported by NSF CAREER Fellowship, DMS 2142064. Iqra Altaf, Ryan Bushling, Bobby Wilson Abstract . We prove that, for every polyhedral or C 1 C^{1} norm on ℝ d \mathbb{R}^{d} and every set E ⊆ ℝ d E\subseteq\mathbb{R}^{d} of packing dimension s s , the packing dimension of the distance set of E E with respect to that norm is at least s d \tfrac{s}{d} . One of the main tools is a nonlinear projection theorem extending a result of M. Järvenpää. An explicit construction follows, demonstrating that these distance sets bounds are sharp for a large class of polyhedral norms. 2020 Mathematics Subject Classification 28A80 1. Introduction For a norm \|\cdot\|_{*} on \mathbb{R}^{d} and set E\subseteq\mathbb{R}^{d} , let \Delta^{*}(E):=\big\{\|y-x\|_{*}\in[0,\infty)\!:x,y\in E\big\} be the distance set of E with respect to that norm. Given a fixed x\in\mathbb{R}^{d} , we also let \Delta_{x}^{*}(E):=\big\{\|y-x\|_{*}\in[0,\infty)\!:y\in E\big\} be the pinned distance set of E at x . The classical Falconer distance conjecture [ 6 ] posits that, for E (say, Borel), \hdim E>\frac{d}{2}\qquad\Longrightarrow\qquad\hdim\Delta(E)=1, where \hdim denotes Hausdorff dimension and \Delta the distance set operator for the Euclidean norm. More quantitatively, one frequently seeks lower bounds on \hdim\Delta(E) or \hdim\Delta_{x}(E) in terms of \hdim E . For recent progress on the distance set conjecture, see [ 5 , 8 , 9 ] . In particular, [ 5 ] contains a clear and thorough history of the problem. This paper may be seen as a follow-up to [ 1 ] , in which the authors treat the Falconer distance problem for general norms, emphasizing in particular the case of polyhedral norms. The main result therein is the following. Theorem ( [ 1 ] Theorems 1.1 & 1.4). Let \|\cdot\|_{*} be a norm on \mathbb{R}^{d} and let E\subseteq\mathbb{R}^{d} . Then (1) \hdim\Delta_{x}^{*}(E)\geq\hdim E-(d-1) for all x\in\mathbb{R}^{d} . This bound is sharp for polyhedral norms in the sense that, if \|\cdot\|_{P} is a polyhedral norm on \mathbb{R}^{d} , then for any s\in[d-1,d] , there exists a compact set E\subset\mathbb{R}^{d} with \hdim E=s such that \hdim\Delta^{P}(E)=s-(d-1). 1.1. Main results Distance sets with respect to polyhedral norms have also been considered in [ 13 , 7 , 12 , 2 ] , and in addition [ 2 ] presents strictly convex C^{1} norms with pathological distance sets. The works [ 14 , 11 , 8 , 9 ] address problems concerning the packing dimension \pdim of distance sets, and here we consider the packing dimension of distance sets with respect to polyhedral and C^{1} norms. Our main theorem is the following. Theorem 1.1. Let \|\cdot\|_{*} be either a C^{1} or polyhedral norm on \mathbb{R}^{d} and let E\subseteq\mathbb{R}^{d} . For every \varepsilon>0 , there exists x\in E such that (2) \pdim\Delta_{x}^{*}(E)\geq\frac{1}{d}\pdim E-\varepsilon. In particular, \pdim\Delta^{*}(E)\geq\frac{1}{d}\pdim E. In both the C^{1} and the polyhedral case, the normal vector field to the surface the unit ball is regular enough to establish reasonable transversality conditions on the pinned distance functions y\mapsto\|x-y\|_{*} . Just as ( 1 ) was sharp for every polyhedral norm in a strong way, ( 2 ) is also sharp for a large class of polyhedral norms. Theorem 1.2. Let \|\cdot\|_{P} be a polyhedral norm such that the unit normal vectors to the faces of the polyhedron are rationally dependent and let s\in[0,d] . There exists a compact set E\subset\mathbb{R}^{d} with \pdim E=s such that (3) \pdim\Delta^{P}(E)=\frac{s}{d}. A main tool in the proof of Theorem 1.1 is a nonlinear projection estimate extending a result of Järvenpää [ 10 ] . Let \mathbf{p}=(\mathbf{p}_{i})_{i=1}^{d} be a family of functions \mathbf{p}_{i}\!:E\to\mathbb{R} defined on a set E\subseteq\mathbb{R}^{d} and let \ubdim denote upper box dimension. Definition 1.1 (Weak transversality). We say that \mathbf{p} is weakly transversal if, for every bounded set F\subseteq E , its restriction \mathbf{p}|_{F} is bounded and has fibers of upper box dimension 0 . That is: i. \mathbf{p}(F) is bounded; and ii. for every \delta\in(0,1) , there exists a natural number M(\delta)\lesssim_{\varepsilon}\delta^{-\varepsilon} for every \varepsilon>0 such that the following holds: for every \xi\in\mathbb{R}^{n} , there exist m\leq M(\delta) many points x_{1},\dots,x_{m}\in\mathbb{R}^{n} such that (4) \mathbf{p}^{-1}\big(B(\xi,\delta)\big)\cap F\subseteq\bigcup_{k=1}^{m}B(x_{k},\delta). The first condition simply ensures that \ubdim\mathbf{p}(F) is defined whenever \ubdim F is. The second condition ensures that the fibers of the functions \mathbf{p}_{i} intersect fairly transversely at most points; it is always satisfied, for example, when \mathbf{p} is locally invertible with Lipschitz inverses. Either \delta (or both) in ( 4 ) can be replaced with a number \sim_{\varepsilon}\delta . With this terminology, we have the following result. Proposition 1.3. Let \mathbf{p}=\big(\mathbf{p}_{i}\!:E\to\mathbb{R}\big)_{i=1}^{d} be a weakly transversal family, let n\in\{1,\dots,d\} , and for every string of indices i_{1},\dots,i_{n} let \mathbf{p}_{i_{1},\dots,i_{n}}:=\big(\mathbf{p}_{i_{j}}\!:E\to\mathbb{R}\big)_{j=1}^{n} . Then for every set F\subseteq E , (5) \frac{n}{d}\ubdim F\leq\max_{1\leq i_{1}<\cdots<i_{n}\leq d}\ubdim\mathbf{p}_{i_{1},\dots,i_{n}}(F) provided F is bounded, and (6) \frac{n}{d}\pdim F\leq\max_{1\leq i_{1}<\cdots<i_{n}\leq d}\pdim\mathbf{p}_{i_{1},\dots,i_{n}}(F). We only require the result for n=1 , and it is possible to prove this directly by a simple covering argument. However, we have included the full statement in the interest of generality and elected a proof that leverages Järvenpää’s result on orthogonal projections. If \|\cdot\|_{*} is a strictly convex norm on \mathbb{R}^{2} , the “projection maps” \mathbf{p}_{i}(x):=\|x-x_{i}\|_{*} are weakly transversal for any two distinct x_{1},x_{2}\in\mathbb{R}^{2} , whereby the distance set bound of Theorem 1.1 follows immediately from Proposition 1.3 . In higher dimensions, one can establish a similar result for strictly convex norms for d pins x_{1},\dots,x_{d} in general position. Things are not so simple for norms with more degenerate curvature. Establishing transversality for some set of pins on a large portion of E becomes the main hurdle in proving distance sets estimates for such norms. 1.2. Outline After § 2 lays out the necessary background on the upper box and packing dimensions, § 3 presents the proof of Proposition 1.3 and § 4 gives the proof of Theorem 1.1 for polyhedral norms. Next, the transversality conditions necessary to apply Proposition 1.3 for projections defined by C^{1} norms are established in § 5 , and in § 6 the proof of Theorem 1.1 for C^{1} norms is completed. Finally, the sharp example of Theorem 1.2 is constructed in § 7 . 2. Preliminaries Let E\subset\mathbb{R}^{d} be any bounded set and N(E,\delta) its covering number at scale \delta —the minimal number of \delta -balls required to cover E . The upper box dimension of a bounded set E\subset\mathbb{R}^{d} is defined by \ubdim E:=\limsup_{\delta\to 0}\frac{\log N(E,\delta)}{-\log\delta}. Note that upper box dimension is finitely stable not but countably stable: that is, if E=\bigcup_{i=1}^{m}E_{i} , then \ubdim E=\max_{1\leq i\leq m}\ubdim E_{i} provided m<\infty , whereas it can happen that \ubdim E>\sup_{1\leq i<\infty}\ubdim E_{i} . Packing dimension can be defined simply by “forcing” upper box dimension to be countably stable: given any set E\subseteq\mathbb{R}^{d} (not necessarily bounded), we let (7) \pdim E:=\inf\left\{\sup_{1\leq i<\infty}\ubdim E_{i}\!:E\subseteq\bigcup_{i=1}^{\infty}E_{i}\right\}, where the infimum is taken over all countable covers of E by bounded sets. It follows immediately that \ubdim E\geq\pdim E for every bounded set E , and it is also not difficult to show that \pdim E\geq\hdim E . We sometimes use the standard notation A\lesssim_{\alpha}B to indicate that A\leq c_{\alpha}\hskip 0.83298ptB for some constant c_{\alpha} depending only on the parameter (or collection of parameters)