Fourier decay and absolute continuity for typical homogeneous self-similar measures in ℝ d {\mathbb{R}}^{d} for d ≥ 3 d\geq 3

Boris Solomyak · 2025

Read Fourier decay and absolute continuity for typical homogeneous self-similar measures in ℝ d {\mathbb{R}}^{d} for d ≥ 3 d\geq 3 by Boris Solomyak (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-2508-14698.txt.

Introduction

Boris Solomyak proved in 2025 new results on the smoothness of fractal measures built by repeated shrinking maps. Such measures arise from iterated function systems, where a few maps copy a set into smaller pieces of itself. The question is whether the measure spreads evenly enough that its wave analysis decays at a power rate.

For systems in three or more dimensions, Solomyak shows the decay holds for nearly every shrink factor, once the pieces span the space. A second result drops the spanning demand, asking only that the pieces not lie flat. The work extends a theory that was well understood only in lower dimensions.

Catalog record

CREATORBoris Solomyak
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE71 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2508-14698

Browse the full library · About the Tower of Babel

Full text

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

Fourier decay and absolute continuity for typical homogeneous self-similar measures in ℝ d {\mathbb{R}}^{d} for d ≥ 3 d\geq 3 Thanks: Supported in part by the Israel Science Foundation grant #1647/23. Boris Solomyak Abstract . We consider iterated function systems (IFS) in ℝ d {\mathbb{R}}^{d} for d ≥ 3 d\geq 3 of the form { f j ( x ) = λ 𝒪 x + a j } j = 0 m \{f_{j}(x)=\lambda{\mathcal{O}}x+a_{j}\}_{j=0}^{m} , with a 0 = 0 a_{0}=0 and m ≥ 1 m\geq 1 . Here λ ∈ ( 0 , 1 ) \lambda\in(0,1) is the contraction ratio and 𝒪 {\mathcal{O}} is an orthogonal matrix. Given a positive probability vector 𝒑 \boldsymbol{p} , there is a unique invariant (stationary) measure for the IFS, called (in this case) a homogeneous self-similar measure, which we denote μ ⁡ ( λ ​ 𝒪 , 𝒟 , 𝒑 ) \mu(\lambda{\mathcal{O}},{\mathcal{D}},\boldsymbol{p}) , where 𝒟 = { a 0 , … , a m } {\mathcal{D}}=\{a_{0},\ldots,a_{m}\} is the set of “vector digits”. We obtain two results on Fourier decay for such measures. First we show that if 𝒟 {\mathcal{D}} spans ℝ d {\mathbb{R}}^{d} , then for every fixed 𝒪 {\mathcal{O}} and 𝒑 \boldsymbol{p} the measure μ ⁡ ( λ ​ 𝒪 , 𝒟 , 𝒑 ) \mu(\lambda{\mathcal{O}},{\mathcal{D}},\boldsymbol{p}) has power Fourier decay (equivalently, positive Fourier dimension) for all but a zero-Hausdorff dimension set of λ \lambda . In our second result we do not impose any restrictions on 𝒟 {\mathcal{D}} , other than the necessary one of affine irreducibility, and obtain power Fourier decay for almost all homogeneous self-similar measures; however, only for even d ≥ 4 d\geq 4 . Combined with recent work of Corso and Shmerkin [ 9 ] , these results imply absolute continuity for almost all self-similar measures under the same assumptions, in the super-critical parameter region. 1. Introduction and statement of results For a finite positive Borel measure \mu on {\mathbb{R}}^{d} , consider the Fourier transform \widehat{\mu}(\xi)=\int_{{\mathbb{R}}^{d}}e^{-2\pi i\langle\xi,x\rangle}\,d\mu(x). The measure is called Rajchman if \widehat{\mu}(\xi) tends to zero at infinity. By the Riemann-Lebesgue Lemma, every absolutely continuous measure is Rajchman, but absolute continuity is not necessary. It is a delicate and much studied question to decide which singular measures are Rajchman. Assuming they are, it is important to obtain specific quantitative bounds for the rate of decay at infinity. Let {\mathcal{M}}({\mathbb{R}}^{d}) be the set of finite positive measures in {\mathbb{R}}^{d} For \gamma>0 consider {\mathscr{D}}_{d}(\gamma)=\bigl\{\nu\in{\mathcal{M}}({\mathbb{R}}^{d}):\ \left|\widehat{\nu}(\xi)\right|=O_{\nu}(|\xi|^{-\gamma}),\ \ |\xi|\to\infty\bigr\}, and denote {\mathscr{D}}_{d}=\bigcup_{\gamma>0}{\mathscr{D}}_{d}(\gamma) . A measure \nu is said to have power Fourier decay if \nu\in{\mathscr{D}}_{d} . This is equivalent to having a positive Fourier dimension , see [ 26 ] . Further, let {\mathscr{D}}_{d,\log}=\bigl\{\nu\in{\mathcal{M}}({\mathbb{R}}^{d}):\ \exists\,\alpha>0,\ \left|\widehat{\nu}(\xi)\right|=O_{\nu}\bigl(\bigl|\log|\xi|\bigr|^{-\alpha}\bigr),\ \ |\xi|\to\infty\bigr\}. Such measures are said to have polylogarithmic Fourier decay . Of course, other decay rates may occur as well. There is on-going very intensive research of the Fourier decay phenomena and its applications, for a broad class of measures which may be termed “fractal”. We are not going to give a comprehensive literature review: a “sample” of recent papers, which includes two surveys, is [ 2 , 3 , 31 , 4 , 5 , 23 ] ; many more references can be found there. However, much of the recent progress is for measures coming from non-linear dynamics, in particular, for stationary (invariant) measure of non-linear iterated function systems (IFS). Here we focus exclusively on the linear case. A measure \mu is called self-affine if it is the invariant measure for a self-affine IFS \{f_{j}\}_{j=0}^{m} , with m\geq 1 , where f_{j}(x)=A_{j}x+a_{j} , the matrices A_{j}:{\mathbb{R}}^{d}\to{\mathbb{R}}^{d} are invertible linear contractions (in some norm) and a_{j}\in{\mathbb{R}}^{d} are “digit” vectors. This means that for some probability vector \boldsymbol{p}=(p_{j})_{j\leq m} holds (1.1) \mu=\sum_{j=0}^{m}p_{j}(\mu\circ f_{j}^{-1}). The IFS is self-similar if all A_{j} are contracting similitudes, that is, A_{j}=\lambda_{j}{\mathcal{O}}_{j} for some \lambda_{j}\in(0,1) and orthogonal matrices {\mathcal{O}}_{j} . The IFS and the corresponding invariant measure are called homogeneous self-affine (respectively, homogeneous self-similar ) if A_{j}=A for all j\leq m , and A is a general linear contracting map (respectively, a contracting similitude). Our results are on self-similar homogeneous IFS, and we review the relevant history with some detail; however, we do not intend to duplicate the excellent survey by Sahlsten [ 31 ] , which is much more exhaustive. Our main motivation for showing the power Fourier decay is to study absolute continuity of self-similar measures, but it has many other applications which we do not discuss here, see, e.g., [ 31 ] . 1.1. Historical background We start with the most basic case: d=1 , m=2 , a homogeneous self-similar measure. Up to an affine-linear conjugacy, we obtain a 1-parameter family of IFS \{\lambda x,\lambda x+1\} . The resulting self-similar measures are known as Bernoulli convolutions ; they were extensively studied since the 1930’s. Denote by \nu_{\lambda}^{p} the Bernoulli convolution, corresponding to the probabilities (p,1-p) . In the “classical case” p=1/2 we omit the superscript p . Erdős [ 12 ] proved that \nu_{\lambda} is not Rajchman when \theta=1/\lambda is a PV (or Pisot) number, that is, an algebraic integer greater than one, all of whose Galois conjugates lie inside the unit circle. Salem [ 32 ] showed that if 1/\lambda is not a Pisot number, then \widehat{\nu}_{\lambda} does vanish at infinity. What about decay rates? Erdős [ 13 ] proved that for any compact interval J\subset(0,1) there exists \alpha=\alpha(J)>0 such that \nu_{\lambda}\in{\mathscr{D}}_{1}(\alpha) for a.e. \lambda\in J . In a short but illuminating expository note, Kahane [ 19 ] indicated that Erdős’ argument actually gives that \nu_{\lambda}\in{\mathscr{D}}_{1} for all \lambda\in(0,1) in the complement of an exceptional set of zero Hausdorff dimension. (This technique became known as the Erdős-Kahane argument , which turned out to be useful in other contexts as well.) All the above results hold for any p\in(0,1) ; moreover, the Erdős-Kahane argument works for any family of homogeneous IFS on the line, with the contraction rate as a parameter. As it often happens with results of “almost sure” kind, few specific \lambda are known, all algebraic, for which \nu_{\lambda} has power Fourier decay. Dai, Feng, and Wang [ 11 ] proved this for the classical Bernoulli convolution \nu_{\lambda} , when \theta=\lambda^{-1} is a Garsia number , i.e., an algebraic integer, all of whose conjugates have modulus greater than 1, and the constant term of its minimal polynomial is \pm 2 . Recently, Streck [ 41 ] generalized this to a larger class of homogeneous IFS, still with a very special algebraic contraction ratio. What about weaker decay rates for explicit \lambda ? Kershner [ 20 ] (see also [ 10 ] ) proved that \nu_{\lambda}\in{\mathscr{D}}_{1,\log} for a rational contraction rate \lambda=p/q , with 1<p<q (note that for \lambda=1/q , with q\geq 3 , the measure \nu_{\lambda} is non-Rajchman; in fact, these are the simplest Pisot Bernoulli convolutions). In a joint work with Bufetov [ 8 , Corollary A.3] it was shown that if \lambda^{-1} is an algebraic integer which has at least one conjugate outside the unit circle, then \nu_{\lambda}\in{\mathscr{D}}_{1,\log} as well; this was extended by Gao and Ma [ 14 ] to an arbitrary set of digits. The case of a Salem number \theta=\lambda^{-1} is very interesting, but remains mysterious. This is, by definition, an algebraic

Read the complete text