betaFeedback

WorksMathematics

Twisted Diophantine approximation: asymptotic and uniform results

Abstract

Fix an m×nm\times n matrix AA; twisted Diophantine approximation studies how well we can approximate the affine form Aq−bA {\bf q}-{\bf b} by an integer mm-vector, depending on the choice of mm-vector b{\bf b}. We report on two companion papers in twisted Diophantine approximation, covering two setups: asymptotic and uniform. The asymptotic paper gives an exact Khintchine-type Lebesgue zero–one law, a null–conull criterion for badly approximable shifts, and Hausdorff-measure tests yielding exact power-law dimensions under a critical divergence condition. The uniform paper computes Hausdorff dimensions of sets of uniformly approximable vectors. It also gives Hausdorff zero–full laws, uniform-exponent spectra, and a sharp escape-of-mass bound. In both cases the results are established for arbitrary approximation functions from suitable regularity classes, and the arguments use a homogeneous lattice count involving all successive minima to record clustering in every independent direction.

Introduction

Diophantine approximation: asymptotic and uniform

Let m,nm,n be positive integers, set d=m+nd=m+n, and write Mat⁡m,n\operatorname{Mat}_{m,n} for the space of real m×nm \times n matrices. Classical theory of homogeneous Diophantine approximation asks, for A∈Mat⁡m,nA \in\operatorname{Mat}_{m,n} and a positive non-increasing function ψ:R+→R+\psi:\mathbb{R}_{+} \to\mathbb{R}_{+}, where R+=(0,∞)\mathbb{R}_{+}=(0,\infty), whether there are integer vectors p∈Zm\mathbf{p} \in\mathbb{Z}^{m} and q∈Zn\mathbf{q} \in\mathbb{Z}^{n} satisfying

∣Aq−p∣≤ψ(T),0<∣q∣≤T(1)|A\mathbf{q}-\mathbf{p}| \le\psi(T), \qquad0 < |\mathbf{q}| \le T \tag*{(1)}

at scale T>0T>0. Here TT bounds the size of q\mathbf{q}, while ψ(T)\psi(T) specifies the allowed approximation error; throughout, ∣⋅∣|\cdot| denotes the supremum norm.

Dirichlet’s theorem (1842) states that (1.1) is solvable for every A∈Mat⁡m,nA \in\operatorname{Mat}_{m,n} and every T>1T>1 when ψ(T)=T−n/m\psi(T)=T^{-n/m}. For a general ψ\psi, we may ask two types of questions: do solutions exist for an unbounded set of TT? Or do solutions exist for all sufficiently large TT? In the former case we call AA ψ\psi-approximable, and in the latter ψ\psi-Dirichlet1; these are the asymptotic and uniform conditions, respectively [20 Section 1.1].

The metric theory asks how large these two classes of matrices are. For the asymptotic condition, the Khintchine–Groshev theorem [11, 7] states that the set of ψ\psi-approximable matrices is null or conull according as

∑q=1∞qn−1ψ(q)m,equivalently,∑k=0∞2knψ(2k)m(2)\sum_{q=1}^{\infty}q^{n-1}\psi(q)^{m},\qquad\text{equivalently,}\qquad\sum_{k=0}^{\infty}2^{kn}\psi(2^{k})^{m} \tag*{(2)}

converges or diverges; see [3 Section 1] for this formulation. For the uniform condition, the set of ψ\psi-Dirichlet matrices is likewise null or conull under suitable regularity assumptions on ψ\psi. This zero–one law was established by Kleinbock and Wadleigh for m=n=1m=n=1 [19], while in arbitrary dimensions the partial result of Kleinbock, Strömbergsson and Yu [18] was completed by Strömbergsson and Yu [27].

Inhomogeneous approximation

In the setup of inhomogeneous Diophantine approximation, we replace the linear form AqA\mathbf{q} by an affine form Aq−bA\mathbf{q}-\mathbf{b}, where b∈Rm\mathbf{b}\in\mathbb{R}^{m}. We study the system

∣Aq−p−b∣≤ψ(T),∣q∣≤T.(3)\lvert A\mathbf{q}-\mathbf{p}-\mathbf{b}\rvert\leq\psi(T),\qquad\lvert\mathbf{q}\rvert\leq T. \tag*{(3)}

The same distinction between asymptotic and uniform approximation now applies to pairs: we call (A,b)(A,\mathbf{b}) ψ\psi-approximable if (3) is solvable for unbounded TT, and ψ\psi-Dirichlet if it is solvable for all sufficiently large TT, with q=0\mathbf{q}=0 allowed here. In the inhomogeneous setting, Dirichlet’s bound is no longer valid for every pair (A,b)(A,\mathbf{b}). Indeed, if AA has integer entries and b∉Zm\mathbf{b}\notin\mathbb{Z}^{m}, then (3) has no solution once ψ(T)\psi(T) is small enough.

The metric results depend on which parameters are allowed to vary. When a nonintegral shift b\mathbf{b} is fixed and AA varies, the asymptotic problem retains the classical measure law: the set of ψ\psi-approximable matrices is null or conull according as the series in (2) converges or diverges [Szű58, Sch64, Spr79]. With both AA and b\mathbf{b} varying, the set of ψ\psi-Dirichlet pairs is also null or conull by the zero–one law of Kleinbock and Wadleigh [20]. The first-named author and Kim further studied the complementary non-Dirichlet sets, establishing Hausdorff-measure laws for pairs and for fixed b∉Zm\mathbf{b}\notin\mathbb{Z}^{m} when ψ\psi is decreasing and tends to zero [14]. In these results, the matrix AA remains a variable. The question addressed in this report is how the two approximation conditions behave when AA is fixed and only the shift b\mathbf{b} varies.

Twisted approximation

This choice is usually called twisted approximation. From now on, we assume ψ(T)→0\psi(T)\to0 as T→∞T\to\infty. Since integer translates of b\mathbf{b} do not affect solvability of (3), shifts naturally belong to Tm=Rm/Zm\mathbb{T}^{m}=\mathbb{R}^{m}/\mathbb{Z}^{m}; we write ⟨x⟩=x+Zm\langle\mathbf{x}\rangle=\mathbf{x}+\mathbb{Z}^{m} for the image of x∈Rm\mathbf{x}\in\mathbb{R}^{m}. For a fixed AA, let WA(ψ)W_A(\psi) and DA(ψ)D_A(\psi) be the sets of shifts for which (3) is solvable for unbounded TT and for all sufficiently large TT, respectively.2

We briefly recall the main metric results for the asymptotic sets WA(ψ)W_A(\psi). Kurzweil’s theorem [21] states that if AA is badly approximable, then WA(ψ)W_A(\psi) has zero or full measure according as the series (2) converges or diverges. Moreover, this law holds for every non-increasing ψ\psi only if AA is badly approximable. Beyond badly approximable matrices, general exact criteria were previously available only in the case m=n=1m=n=1: Fuchs and Kim [6] gave an exact criterion based on the continued fraction expansion; later, Kim, Rams and Wang [17] determined the exact Hausdorff dimensions of these sets. For nonsingular matrices, the first-named author [12] gave a sufficient full-measure condition, extended to the weighted setting by Hussain and Ward [9]. The asymptotic problem we consider is to determine when WA(ψ)W_A(\psi) is null or conull for an arbitrary fixed AA.

We next recall the main metric results for the uniform sets DA(ψ)D_A(\psi). Transference relates the approximation of shifts to homogeneous approximation of ATA^{\mathsf{T}}. Jarník [Jar39] gave conditions for every shift to be uniformly approximable, proving an analogue of Dirichlet’s theorem in the twisted setup. Bugeaud and Laurent [4] related almost-everywhere asymptotic exponents to uniform homogeneous exponents, and Moshchevitin and the second-named author [22] obtained a zero–one law for the Lebesgue measure of DA(ψ)D_A(\psi), up to multiplicative constants in ψ\psi. These results do not in general determine the Hausdorff dimension of DA(ψ)D_A(\psi) when it is Lebesgue null. Kim and Liao [13] obtained Hausdorff-dimension results for irrational rotations when ψ\psi is a power function. Aggarwal [1] obtained dimension bounds for singular affine forms. The uniform problem we consider is to determine dimensions for arbitrary fixed matrices.

The two companion papers

The two companion papers address these questions for arbitrary fixed matrices AA. For non-increasing ψ\psi satisfying the regularity condition (6), the asymptotic paper [15] gives a convergence–divergence criterion deciding whether WA(ψ)W_A(\psi) is null or conull. It also derives results on twisted badly approximable vectors and Hausdorff measure and dimension.

For continuous, strictly decreasing ψ\psi satisfying the regularity condition (8), the uniform paper [16] computes the Hausdorff dimensions of the sets ⋂c>0DA(cψ)\bigcap_{c>0}D_A(c\psi) and ⋃c>0DA(cψ)\bigcup_{c>0}D_A(c\psi), and in many cases of DA(ψ)D_A(\psi). For powers and near powers, it determines dim⁡HDA(cψ)\dim_{\mathrm{H}}D_A(c\psi) for every c>0c>0, except at finitely many exponents depending only on m,nm,n.

Different denominator vectors can give the same or nearby points in the projected orbit {⟨Aq⟩:q∈Zn}\{\langle A\mathbf{q}\rangle:\mathbf{q}\in\mathbb{Z}^n\} when AA maps their differences close to integer vectors. Both papers express their formulae through a homogeneous lattice count that measures this clustering. § 2 states the precise hypotheses and main results, and § 3 explains how this count enters the two arguments. Complete proofs are given in the companion papers.

Conventions. All balls in Tm\mathbb{T}^{m} use the quotient supremum metric, and dim⁡H\dim_{H} denotes Hausdorff dimension. Write Lm\mathcal{L}^{m} for normalized Lebesgue measure on Tm\mathbb{T}^{m}. The symbols ≪\ll, ≫\gg and ≍\asymp allow positive constants depending on m,nm,n, unless additional fixed parameters are indicated.

Use of artificial intelligence. ChatGPT Codex was used in editing, proofreading and creating figures for this article. The authors take full responsibility for all the content of this manuscript.

Acknowledgements. The authors are grateful to Dmitry Kleinbock for useful suggestions and discussion.

Main results

First, we introduce the lattice count which enters both asymptotic and uniform results.

Lattice counts and the diagonal action

To encode the Diophantine properties of AA, associate to it the unimodular lattice

ΛA:=(ImA0In)Zd={(Aq−p,q):(p,q)∈Zm×Zn}.(4)\Lambda_A := \begin{pmatrix} I_m & A \\ 0 & I_n \end{pmatrix}\mathbb{Z}^d = \{(Aq-p,q):(p,q)\in\mathbb{Z}^m\times\mathbb{Z}^n\}. \tag*{(4)}

For r,T>0r,T>0, define

NA(r,T):=#(ΛA∩([−r,r]m×[−T,T]n)).N_A(r,T):=\#\left(\Lambda_A\cap\left([-r,r]^m\times[-T,T]^n\right)\right).

Thus NA(r,T)N_A(r,T) counts integer pairs with error at most rr and denominator size at most TT. Such pairs describe differences between nearby projected orbit points, so the count records their clustering. It includes the zero vector and is therefore always positive.

We use successive minima relative to the fixed unit box B=[−1,1]dB=[-1,1]^d. For a full-rank lattice Λ\Lambda, let λi(Λ)\lambda_i(\Lambda) be the least λ>0\lambda>0 such that λB\lambda B contains ii linearly independent vectors of Λ\Lambda. Consider the diagonal action

at:=(et/mIm00e−t/nIn),t∈R.a_t:=\begin{pmatrix} e^{t/m}I_m & 0 \\ 0 & e^{-t/n}I_n \end{pmatrix},\qquad t\in\mathbb{R}.

Write λi(t):=λi(atΛA)\lambda_i(t):=\lambda_i(a_t\Lambda_A). For r,T>0r,T>0, put

Rr,T:=(rmTn)1/d,tr,T:=mndlog⁡Tr.R_{r,T}:=(r^mT^n)^{1/d},\qquad t_{r,T}:=\frac{mn}{d}\log\frac{T}{r}.

The map atr,Ta_{t_{r,T}} sends the rectangle [−r,r]m×[−T,T]n[-r,r]^m\times[-T,T]^n to Rr,TBR_{r,T}B, so NA(r,T)=#(atr,TΛA∩Rr,TB)N_A(r,T)=\#(a_{t_{r,T}}\Lambda_A\cap R_{r,T}B). Standard lattice-counting estimates therefore give

NA(r,T)≍∏i=1dmax⁡{1,Rr,Tλi(tr,T)}.(5)N_A(r,T)\asymp\prod_{i=1}^{d}\max\left\{1,\frac{R_{r,T}}{\lambda_i(t_{r,T})}\right\}. \tag*{(5)}

This comparison is uniform in A,r,TA,r,T [8, 15]. Thus the count is determined by the successive minima along atΛAa_t\Lambda_A, connecting both approximation problems with parametric geometry of numbers [25, 5].

A zero–one law for asymptotic approximation

We call ψ\psi dyadically regular if, for some 0<ϱ<10 < \varrho< 1,

ψ(2k+1)≤ϱψ(2k)for all sufficiently large k.(6)\psi(2^{k+1}) \le\varrho\psi(2^k) \qquad\text{for all sufficiently large } k. \tag*{(6)}

For instance, if v>0v > 0, the power function ψv(T):=T−v\psi_v(T) := T^{-v} satisfies this condition. The asymptotic measure law is as follows.

Theorem 2.1 [15 Theorem 1.1]. Let A∈Mat⁡m,nA \in\operatorname{Mat}_{m,n} and let ψ:R+→R+\psi: \mathbb{R}_+ \to\mathbb{R}_+ be non-increasing and satisfy (6). Then

Lm(WA(ψ))={0,if ∑k=0∞2knψ(2k)mNA(ψ(2k),2k)<∞,1,if ∑k=0∞2knψ(2k)mNA(ψ(2k),2k)=∞.(7)\mathcal{L}^{m}(W_A(\psi)) = \begin{cases} 0, & \text{if } \sum_{k=0}^{\infty}\frac{2^{kn}\psi(2^k)^m}{N_A(\psi(2^k),2^k)} < \infty,\\[1.2em] 1, & \text{if } \sum_{k=0}^{\infty}\frac{2^{kn}\psi(2^k)^m}{N_A(\psi(2^k),2^k)} = \infty. \end{cases} \tag*{(7)}

The summand in (7) separates volume from clustering:

2knψ(2k)mNA(ψ(2k),2k)=Khintchine–Groshev block volumehomogeneous clustering multiplicity.\frac{2^{kn}\psi(2^k)^m}{N_A(\psi(2^k),2^k)} = \frac{\text{Khintchine--Groshev block volume}}{\text{homogeneous clustering multiplicity}}.

The numerator is the kk-th dyadic term in (2), while the denominator corrects for clustering in the projected orbit. For badly approximable matrices and irrational rotations, this recovers the Kurzweil and Fuchs–Kim criteria, respectively, within the class of dyadically regular functions.

Lagrange constants and badly approximable shifts

Define the inhomogeneous ψ\psi-Lagrange constant by

cA,ψ(b):=lim inf⁡∣q∣→∞min⁡p∈Zm∣Aq−p−b∣ψ(∣q∣).c_{A,\psi}(\mathbf{b}) := \liminf_{\lvert\mathbf{q}\rvert\to\infty} \frac{\min_{\mathbf{p}\in\mathbb{Z}^{m}}\lvert A\mathbf{q}-\mathbf{p}-\mathbf{b}\rvert}{\psi(\lvert\mathbf{q}\rvert)}.

Under the hypotheses of Theorem 2.1, this constant is almost surely 00 or +∞+\infty according as the series in (7) diverges or converges [15 Corollary 2.1]. At the Dirichlet exponent n/mn/m, define Bad⁡A:={b∈Tm:cA,ψn/m(b)>0}\operatorname{Bad}_A := \{\mathbf{b} \in\mathbb{T}^m : c_{A,\psi_{n/m}}(\mathbf{b}) > 0\}. At time t=nklog⁡2t = nk\log2, (5) gives

Δk(A):=∏i=1dmin⁡{1,λi(anklog⁡2ΛA)}≍NA(2−kn/m,2k)−1.\Delta_k(A) := \prod_{i=1}^{d}\min\{1,\lambda_i(a_{nk\log2}\Lambda_A)\}\asymp N_A(2^{-kn/m},2^k)^{-1}.

The measure of badly approximable shifts is then determined as follows.

Theorem 2.2 [15 Proposition 2.2]. Let A∈Mat⁡m,nA \in\operatorname{Mat}_{m,n}. Then

Lm(Bad⁡A)={0,if ∑kΔk(A)=∞,1,if ∑kΔk(A)<∞.\mathcal{L}^{m}(\operatorname{Bad}_A) = \begin{cases} 0, & \text{if } \sum_k \Delta_k(A)=\infty,\\[0.8em] 1, & \text{if } \sum_k \Delta_k(A)<\infty. \end{cases}

This answers the question in [12 Remark 1.8(3)] and the Lebesgue-measure case of [MRS24, Question 1.3].

Recall that AA is singular if, for every ε>0\varepsilon> 0, (1) is solvable for all sufficiently large TT with ψ(T)=εT−n/m\psi(T) = \varepsilon T^{-n/m}; otherwise it is nonsingular. For a nonsingular matrix, the count NA(2−kn/m,2k)N_A(2^{-kn/m},2^k) is bounded for infinitely many kk, so ∑kΔk(A)=∞\sum_k \Delta_k(A) = \infty and Bad⁡A\operatorname{Bad}_A is null [15 Corollary 2.4].

Asymptotic Hausdorff measure and dimension

To describe the size of asymptotic approximation sets more finely, we turn to Hausdorff measure. For non-increasing, dyadically regular ψ\psi and 0<s<m0 < s < m, the count gives two sufficient conditions [15 Corollary 3.1]:

Hs(WA(ψ))={0,if ∑k=0∞2knψ(2k)sNA(ψ(2k),2k)<∞,∞,if ∑k=0∞2knψ(2k)sNA(ψ(2k)s/m,2k)=∞.\mathcal{H}^{s}(W_A(\psi)) = \begin{cases} 0, & \text{if } \sum_{k=0}^{\infty}\frac{2^{kn}\psi(2^k)^s}{N_A(\psi(2^k),2^k)} < \infty,\\ \infty, & \text{if } \sum_{k=0}^{\infty}\frac{2^{kn}\psi(2^k)^s}{N_A(\psi(2^k)^{s/m},2^k)} = \infty. \end{cases}

The two denominators need not agree, so these implications do not form a single convergence–divergence criterion. The first counts clusters at the original radius, while the second uses the enlarged radius ψ(2k)s/m\psi(2^k)^{s/m} required by mass transference. When the critical series ∑kΔk(A)\sum_k \Delta_k(A) diverges, these tests nevertheless determine the Hausdorff dimension for every power function [15 Corollary 3.2]:

∑kΔk(A)=∞⟹dim⁡HWA(ψv)=min⁡{m,nv}(v>0).\sum_k \Delta_k(A) = \infty\quad\Longrightarrow\quad\dim_{\mathrm{H}} W_A(\psi_v) = \min\left\{m,\frac{n}{v}\right\}\quad(v>0).

In particular, this formula holds for every nonsingular AA.

Uniform Hausdorff dimension

The dimension of DA(cψ)D_A(c\psi) may depend on cc. To eliminate this dependence, define the endpoint sets

Sing⁡A(ψ):=⋂c>0DA(cψ),UA(ψ):=⋃c>0DA(cψ).\operatorname{Sing}_A(\psi) := \bigcap_{c>0}D_A(c\psi),\qquad U_A(\psi) := \bigcup_{c>0}D_A(c\psi).

We refer to the elements of Sing⁡A(ψn/m)\operatorname{Sing}_A(\psi_{n/m}) simply as twisted singular vectors; see [1 Section 2.4] for the corresponding fixed-matrix problem for singular affine forms.

We call ψ\psi lower quasimultiplicative if there are τ>0\tau> 0 and R0>1R_0 > 1 such that

R−τψ(T)≤ψ(RT)for all large T and every R≥R0.(8)R^{-\tau}\psi(T) \le\psi(RT)\qquad\text{for all large }T\text{ and every }R\ge R_0. \tag*{(8)}

For continuous, strictly decreasing ψ\psi tending to zero, we will assume ψ(T)→∞\psi(T)\to\infty as T→0T\to0, so that ψ−1\psi^{-1} is defined on R+\mathbb{R}_+. This assumption does not affect the set DA(ψ)D_A(\psi).

For L,c>0L,c>0, set

Tj:=ψ−1(e−jL),j≥0.(9)T_j := \psi^{-1}(e^{-jL}),\qquad j\ge0. \tag*{(9)}

and

PA,ψ(L,c):=lim inf⁡N→∞1NL∑j=0N−1log⁡NA(cψ(Tj),Tj+1)NA(cψ(Tj+1),Tj+1).(10)P_{A,\psi}(L,c):=\liminf_{N\to\infty}\frac{1}{NL}\sum_{j=0}^{N-1}\log\frac{N_A(c\psi(T_j),T_{j+1})}{N_A(c\psi(T_{j+1}),T_{j+1})}. \tag*{(10)}

The count ratio governs how many smaller approximation balls can be retained at each step; NL=−log⁡ψ(TN)NL=-\log\psi(T_N) measures the total logarithmic shrinkage of the radii. Our main result regarding the uniform setup is as follows.

Theorem 2.3 [16 Theorem 1.2]. Let A∈Mat⁡m,nA\in\operatorname{Mat}_{m,n} and let ψ\psi be continuous, strictly decreasing, lower quasimultiplicative and tend to zero.

(i) The limits

P−(A,ψ):=lim inf⁡L→∞c>0PA,ψ(L,c),P_-(A,\psi):=\liminf_{\substack{L\to\infty\\c>0}}P_{A,\psi}(L,c),
P+(A,ψ):=lim sup⁡L→∞c>0PA,ψ(L,c)P_+(A,\psi):=\limsup_{\substack{L\to\infty\\c>0}}P_{A,\psi}(L,c)

exist, and

dim⁡HSing⁡A(ψ)=inf⁡c>0dim⁡HDA(cψ)=P−(A,ψ),\dim_H\operatorname{Sing}_A(\psi)=\inf_{c>0}\dim_H D_A(c\psi)=P_-(A,\psi),
dim⁡HUA(ψ)=sup⁡c>0dim⁡HDA(cψ)=P+(A,ψ).\dim_H U_A(\psi)=\sup_{c>0}\dim_H D_A(c\psi)=P_+(A,\psi).

(ii) The quantity dim⁡HDA(cψ)\dim_H D_A(c\psi) is independent of c>0c>0 if and only if P−(A,ψ)=P+(A,ψ)P_-(A,\psi)=P_+(A,\psi). Under this condition,

dim⁡HDA(cψ)=lim⁡L→∞PA,ψ(L,1)for every c>0.\dim_H D_A(c\psi)=\lim_{L\to\infty}P_{A,\psi}(L,1)\qquad\text{for every }c>0.

In particular, dim⁡HDA(ψ)\dim_H D_A(\psi) lies between P−P_- and P+P_+.

Under the hypotheses of Theorem 2.3, both Sing⁡A(ψ)\operatorname{Sing}_A(\psi) and UA(ψ)U_A(\psi) are Borel subgroups of HA:={⟨Aq⟩:q∈Zn}H_A:=\{\langle Aq\rangle:q\in\mathbb{Z}^n\} and each has either zero or locally full Hf\mathcal{H}^f-measure for every dimension function ff [16 Theorem 1.9]. Here f:R+→R+f:\mathbb{R}_+\to\mathbb{R}_+ is continuous and nondecreasing with f(r)→0f(r)\to0 as r→0r\to0; locally full means Hf(E∩B)=Hf(B)\mathcal{H}^f(E\cap B)=\mathcal{H}^f(B) for every nonempty relative ball BB in HAH_A.

Power functions and uniform exponents

For power functions, the dimension of DA(cψτ)D_A(c\psi_\tau) is independent of c>0c>0 outside the finite set

Em,n:={ba:a,b∈N, 1≤a≤m, 1≤b≤n}.\mathcal{E}_{m,n}:=\left\{\frac{b}{a}:a,b\in\mathbb{N},\ 1\leq a\leq m,\ 1\leq b\leq n\right\}.

In this case, we can state the dimension formula in a simpler way than in Theorem 2.3. Put

tτ(T):=mn(1+τ)dlog⁡T,rA,τ(T):=#{1≤i≤d:λi(tτ(T))<T(n−mτ)/d}.t_\tau(T):=\frac{mn(1+\tau)}{d}\log T,\qquad r_{A,\tau}(T):=\#\{1\leq i\leq d:\lambda_i(t_\tau(T))<T^{(n-m\tau)/d}\}.

Thus rA,τ(T)r_{A,\tau}(T) counts the minima below T(n−mτ)/dT^{(n-m\tau)/d} at time tτ(T)t_\tau(T) along $a_t\Lambda_A.`

Theorem 2.4 [16 Theorem 1.4]. Let A∈Mat⁡m,nA \in\operatorname{Mat}_{m,n} and τ∈(0,∞)∖Em,n\tau\in(0,\infty) \setminus\mathcal{E}_{m,n}. For every c>0c>0,

dim⁡HDA(cψτ)=lim inf⁡T→∞∫1TrA,τ(S)dSS−log⁡NA(T−τ,T)(1+τ)log⁡T.(11)\dim_{H} D_A(c\psi_\tau)=\liminf_{T\to\infty}\frac{\int_1^T r_{A,\tau}(S)\frac{\mathrm{d}S}{S}-\log N_A(T^{-\tau},T)}{(1+\tau)\log T}. \tag*{(11)}

Write δA(τ)\delta_A(\tau) for this common value. It is continuous on each component of (0,∞)∖Em,n(0,\infty)\setminus\mathcal{E}_{m,n}. Moreover, if ψ\psi is continuous, strictly decreasing to zero, and

lim⁡T→∞log⁡ψ(T)−1log⁡T=τ,\lim_{T\to\infty}\frac{\log\psi(T)^{-1}}{\log T}=\tau,

then

dim⁡HDA(cψ)=δA(τ)for every c>0.\dim_H D_A(c\psi)=\delta_A(\tau)\qquad\text{for every }c>0.

No lower quasimultiplicativity is required for this last assertion. For these powers, both Sing⁡A(ψτ)\operatorname{Sing}_A(\psi_\tau) and UA(ψτ)U_A(\psi_\tau) have dimension δA(τ)\delta_A(\tau). The near-power assertion applies, for example, to ψ(T)=T−τ(log⁡T)β\psi(T)=T^{-\tau}(\log T)^\beta for every fixed β∈R\beta\in\mathbb{R}.

The exclusions are necessary: for each τ=b/a∈Em,n\tau=b/a\in\mathcal{E}_{m,n}, there is a matrix AA with dim⁡HDA(cψτ)=0\dim_H D_A(c\psi_\tau)=0 for small cc and dim⁡HDA(cψτ)=a\dim_H D_A(c\psi_\tau)=a for large cc [16 Remark 1.5]. In the m=n=1m=n=1 case, for irrational AA, continued fractions give formulae for dim⁡HDA(ψ)\dim_H D_A(\psi) under the hypotheses of Theorem 2.3 whenever the ratio log⁡ψ(T)−1/log⁡T\log\psi(T)^{-1}/\log T stays uniformly below or above 11 [16 Theorem 2.2], extending [13].

These formulae also describe the uniform exponent, defined by ω^(A,b):=sup⁡{s>0:b∈DA(ψs)}\widehat{\omega}(A,\mathbf{b}):=\sup\{s>0:\mathbf{b}\in D_A(\psi_s)\}, with sup⁡∅=0\sup\varnothing=0. Write δA(τ±)\delta_A(\tau^\pm) for the one-sided limits of the non-increasing function δA\delta_A, taken through exponents outside Em,n\mathcal{E}_{m,n}. For every τ>0\tau>0, including those belonging to the exceptional set Em,n\mathcal{E}_{m,n}, [16 Proposition 1.6] gives

dim⁡H{b:ω^(A,b)≥τ}=δA(τ−),\dim_H\{\mathbf{b}:\widehat{\omega}(A,\mathbf{b})\geq\tau\}=\delta_A(\tau^-),
dim⁡H{b:ω^(A,b)>τ}=δA(τ+).\dim_H\{\mathbf{b}:\widehat{\omega}(A,\mathbf{b})>\tau\}=\delta_A(\tau^+).

Every nonempty exact level {b:ω^(A,b)=τ}\{\mathbf{b}:\widehat{\omega}(A,\mathbf{b})=\tau\} has dimension δA(τ−)\delta_A(\tau^-).

Twisted singular vectors

2.7. Twisted singular vectors. For singular affine forms, a simpler bound uses only the first minimum along atΛAa_t\Lambda_A. Define the lower escape of mass by

Emass⁡(A):=lim⁡ε↓0lim inf⁡X→∞1X∫0X1{λ1(t)<ε} dt.\operatorname{Emass}(A):=\lim_{\varepsilon\downarrow0}\liminf_{X\to\infty}\frac{1}{X}\int_0^X\mathbf{1}_{\{\lambda_1(t)<\varepsilon\}}\,\mathrm{d}t.

It measures the lower proportion of time spent with an arbitrarily small first minimum.

Theorem 2.5 [16 Propositions 2.4–2.5]. For every A∈Mat⁡m,nA\in\operatorname{Mat}_{m,n},

dim⁡HSing⁡A(ψn/m)≤m(d−1)dEmass⁡(A).\dim_H\operatorname{Sing}_A(\psi_{n/m})\leq\frac{m(d-1)}{d}\operatorname{Emass}(A).

For every η∈[0,1]\eta\in[0,1], there is an A∈Mat⁡m,nA \in\operatorname{Mat}_{m,n} with Emass⁡(A)=η\operatorname{Emass}(A)=\eta attaining equality.

This improves the coefficient mm in [1]; the theorem is optimal when only lower escape of mass is used.

The common geometry: clusters and scales

Fix a large TT. Partition Tm\mathbb{T}^{m} into half-open cubes of side comparable to ψ(T)\psi(T) and at most ψ(T)\psi(T); in the supremum norm these are ball-shaped cells. Denote this partition by P(T)\mathcal{P}(T), and call a cell good if it contains at least one orbit point with denominator bound TT:

G(T):={B∈P(T):⟨Aq⟩∈B for some ∣q∣≤T},E(T):=⋃B∈G(T)B.\mathcal{G}(T):=\{B\in\mathcal{P}(T):\langle A\mathbf{q}\rangle\in B\text{ for some }|\mathbf{q}|\le T\},\qquad E(T):=\bigcup_{B\in\mathcal{G}(T)}B.

Every shift in a good cell satisfies (3), since it is within ψ(T)\psi(T) of the orbit point in that cell. Conversely, a shift satisfying (3) lies within distance ψ(T)\psi(T) of a good cell. Thus the partition description agrees with the approximation condition up to a multiplicative change of radius, bounded independently of AA and TT. Figure 1 shows the question we need to answer: how many cells are good?

Partition $\mathcal{P}(T)$ and good cells $\mathcal{G}(T)$, with blue orbit points $\langle A\mathbf{q}\rangle$, $|\mathbf{q}|\le T$, orange good cells, and the number of points in a cluster $N_A(\psi(T),T)$

Figure 1. Blue orbit centers visit the orange good cells. The partition P(T)\mathcal{P}(T) has cells of side ≍ψ(T)\asymp\psi(T); the total number of points is ≍Tn\asymp T^n, the number of points in one cell is NA(ψ(T),T)N_A(\psi(T),T), and #G(T)≍total number of pointsnumber of points in one cell≍TnNA(ψ(T),T)\#\mathcal{G}(T)\asymp\frac{\text{total number of points}}{\text{number of points in one cell}}\asymp\frac{T^n}{N_A(\psi(T),T)}.

There are about TnT^n points ⟨Aq⟩\langle A\mathbf{q}\rangle with ∣q∣≤T|\mathbf{q}|\le T, counted with multiplicity, but this need not give TnT^n different good cells. Several vectors can visit the same cell, or even give exactly the same torus point. To understand this multiplicity, lift the picture from Tm\mathbb{T}^{m} to the lattice ΛA⊂Rm+n\Lambda_A\subset\mathbb{R}^{m+n} in (4). The extra coordinates keep track of the denominator vectors q\mathbf{q} rather than forgetting them under projection.

First count all visits. For each q∈Zn\mathbf{q} \in\mathbb{Z}^{n} with ∣q∣≤T|\mathbf{q}| \le T, there is exactly one p∈Zm\mathbf{p} \in\mathbb{Z}^{m} for which Aq−p∈[0,1)mA\mathbf{q}-\mathbf{p} \in[0,1)^{m}. Hence

#(ΛA∩([0,1)m×[−T,T]n))=(2⌊T⌋+1)n≍Tn.\#\left(\Lambda_A \cap\left([0,1)^{m} \times[-T,T]^{n}\right)\right) = (2\lfloor T\rfloor+1)^{n} \asymp T^{n}.

This is comparable to the volume of the cylinder.

Next look at the visits to a good cell BB. In a lift of BB, these correspond to the points of ΛA∩(B×[−T,T]n)\Lambda_A \cap\left(B \times[-T,T]^{n}\right). Choose one such point (Aq0−p0,q0)(A\mathbf{q}_0-\mathbf{p}_0,\mathbf{q}_0). Since ΛA\Lambda_A is a lattice, the translation which moves this point to the origin preserves ΛA\Lambda_A. The number of points in a good cell is then measured, up to fixed enlargements of the window B×[−T,T]nB \times[-T,T]^{n}, by NA(ψ(T),T)N_A(\psi(T),T) (Figure 2).

Lattice window and orbit clusters

Figure 2. The orange window has base centered at x0\mathbf{x}_0. Lattice points in the orange window project to a cluster of orbit centers (m=2m=2, n=1n=1).

Fixed enlargements are needed because the chosen lattice point λ0=(Aq0−p0,q0)\lambda_0=(A\mathbf{q}_0-\mathbf{p}_0,\mathbf{q}_0) need not be the center of its window. If BB has side length ℓ≍ψ(T)\ell\asymp\psi(T), then (3B×[−2T,2T]n)−λ0(3B \times[-2T,2T]^{n})-\lambda_0 contains [−ℓ,ℓ]m×[−T,T]n[-\ell,\ell]^{m} \times[-T,T]^{n}, where 3B3B is the concentric triple dilate of BB. The enlarged window therefore contains at least NA(ℓ,T)≍NA(ψ(T),T)N_A(\ell,T) \asymp N_A(\psi(T),T) lattice points. These changes cost only fixed factors, since (5) gives

NA(c1r,c2T)≍c1,c2NA(r,T)(c1,c2>0).N_A(c_1r,c_2T) \asymp_{c_1,c_2} N_A(r,T) \qquad(c_1,c_2>0).

One short lattice direction gives strings of nearby visits; several short directions give higher-dimensional clusters. This is why the count involves all successive minima, rather than only the shortest vector.

We can now read off the one-scale estimate as

#G(T)≍total number of pointsnumber of points in one cell≍TnNA(ψ(T),T).\#\mathcal{G}(T) \asymp\frac{\text{total number of points}}{\text{number of points in one cell}} \asymp\frac{T^{n}}{N_A(\psi(T),T)}.

Each cell has volume comparable to ψ(T)m\psi(T)^m, so

Lm(E(T))≍ψ(T)m#G(T)≍ψ(T)mTnNA(ψ(T),T).\mathcal{L}^{m}(E(T)) \asymp\psi(T)^m \#\mathcal{G}(T) \asymp\frac{\psi(T)^m T^n}{N_A(\psi(T),T)}.

Taking T=2kT=2^k gives exactly the summand in Theorem 2.1: the usual block volume, divided by its clustering multiplicity. The same volume estimate holds for the cover by enlarged cells. If these volumes are summable, Borel–Cantelli gives measure zero. For divergence, we must also understand how the good cells at different scales overlap.

Fix S<TS<T and an earlier good cell BS∈G(S)B_S\in\mathcal{G}(S). At the later time TT, the cells have diameter comparable to ψ(T)\psi(T). We ask how many cells BT∈G(T)B_T\in\mathcal{G}(T) meet BSB_S. The same total/local picture now takes place inside the parent cylinder:

ΛA∩(BS×[−T,T]n)⏟total visits in the parent,ΛA∩(BT×[−T,T]n)⏟local visits in a finer cell.\underbrace{\Lambda_A\cap\left(B_S\times[-T,T]^n\right)}_{\text{total visits in the parent}}, \qquad \underbrace{\Lambda_A\cap\left(B_T\times[-T,T]^n\right)}_{\text{local visits in a finer cell}}.

To count the TT-level cells intersecting an SS-level cell, center the parent BSB_S at one of its orbit points and give it radius ψ(S)\psi(S). For the lower bound, we also assume T≥2ST\ge2S and ψ(T)≤ψ(S)/4\psi(T)\le\psi(S)/4. Translation by this lattice point gives

number of level-T good cells contained in BSnumber of points in one level-S cell≍number of points in one level-S cellnumber of points in one level-T cell≍NA(ψ(S),T)NA(ψ(T),T).(12)\frac{\text{number of level-}T\text{ good cells contained in }B_S} {\text{number of points in one level-}S\text{ cell}} \asymp \frac{\text{number of points in one level-}S\text{ cell}} {\text{number of points in one level-}T\text{ cell}} \asymp \frac{N_A(\psi(S),T)}{N_A(\psi(T),T)}. \tag*{(12)}

The upper bound for cells meeting the original parent holds for every S<TS<T. Both counts use the later denominator bound TT (Figure 3).

Diagram showing points in $B_S$ and good cells at level $T$, with labels “Points in $B_S$”, “Good cells at level $T$”, “Total $\approx N_A(\psi(S),T)$”, “Local $\approx N_A(\psi(T),T)$”, and “number of level-$T$ good cells $\approx \frac{\text{total}}{\text{local}}=\frac{N_A(\psi(S),T)}{N_A(\psi(T),T)}$”

Figure 3. Inside BSB_S, later orbit centers are grouped at scale ψ(T)\psi(T).

The asymptotic problem. The convergence case follows from the volume estimate and Borel–Cantelli, as shown above. For divergence, put Tk=2kT_k=2^k, Ek=E(Tk)E_k=E(T_k), and

Mk:=ψ(Tk)mTknNA(ψ(Tk),Tk),SN:=∑k=k0NMk.M_k:=\frac{\psi(T_k)^mT_k^n}{N_A(\psi(T_k),T_k)},\qquad S_N:=\sum_{k=k_0}^{N}M_k.

where k0k_0 is chosen so that ψ(Tk+1)≤ϱψ(Tk)\psi(T_{k+1})\le\varrho\psi(T_k) for k≥k0k\ge k_0, with 0<ϱ<10<\varrho<1. One has SN≍∑k=k0NLm(Ek)S_N\asymp\sum_{k=k_0}^{N}\mathcal{L}^m(E_k). We seek the following bound on total pairwise overlaps, which would show quasi-independence of the sets EkE_k on average:

∑k,ℓ=k0NLm(Ek∩Eℓ)≪SN2+SN.(13)\sum_{k,\ell=k_0}^{N}\mathcal{L}^m(E_k\cap E_\ell)\ll S_N^2+S_N. \tag*{(13)}

Suppose SN→∞S_N\to\infty. A generalization of the divergent part of the Borel–Cantelli lemma, known as the Kochen–Stone inequality (see [26 Lemma 5]), yields that Lm(WA(ψ))>0\mathcal{L}^m(W_A(\psi))>0. A separate statement proved in the asymptotic paper shows that this measure is either zero or one; hence it equals one.

It remains to prove the quasi-independence estimate (13). Counting later good cells near each earlier good cell gives

Lm(Ek∩Eℓ)≪TknNA(ψ(Tk),Tk)⏟number of earlier good cellsNA(ψ(Tk),Tℓ)NA(ψ(Tℓ),Tℓ)⏟later good cells meeting one earlier cellψ(Tℓ)m⏟volume of one later cell=MkMℓNA(ψ(Tk),Tℓ)ψ(Tk)mTℓn(k<ℓ).\begin{aligned} \mathcal{L}^m(E_k\cap E_\ell)\ll \underbrace{\frac{T_k^n}{N_A(\psi(T_k),T_k)}}_{\text{number of earlier good cells}} \underbrace{\frac{N_A(\psi(T_k),T_\ell)}{N_A(\psi(T_\ell),T_\ell)}}_{\text{later good cells meeting one earlier cell}} \underbrace{\psi(T_\ell)^m}_{\text{volume of one later cell}} \\ &=M_kM_\ell\frac{N_A(\psi(T_k),T_\ell)}{\psi(T_k)^mT_\ell^n}\qquad(k<\ell). \end{aligned}

We split the pairs k<ℓk<\ell according to whether

ΛA∩([−ψ(Tk),ψ(Tk)]m×[−Tℓ,Tℓ]n)\Lambda_A\cap\left([-\psi(T_k),\psi(T_k)]^m\times[-T_\ell,T_\ell]^n\right)

spans Rm+n\mathbb{R}^{m+n}. In the full-rank case, (5) shows that the count is comparable to the window volume, giving a total overlap contribution O(SN2)O(S_N^2). In case of any smaller rank, the estimates in [15], using the dyadic regularity of ψ\psi, give a total contribution Oϱ(SN)O_\varrho(S_N). We add the diagonal contribution O(SN)O(S_N) and prove (13).

The uniform problem. We use the full approximation targets

Fc(T):=⋃q∈Zn∣q∣≤TB(⟨Aq⟩,cψ(T)),c>0.F_c(T):=\bigcup_{\substack{\mathbf q\in\mathbb{Z}^n\\|\mathbf q|\le T}}B(\langle A\mathbf q\rangle,c\psi(T)),\qquad c>0.

Fix cc and choose the levels TjT_j from (9), so ψ(Tj)=e−jL\psi(T_j)=e^{-jL}. For large LL, lower quasimultiplicativity ensures Tj+1≥2TjT_{j+1}\ge2T_j for all sufficiently large jj.

Starting at such an index kk, (12) gives an inner construction with disjoint children inside each parent and an outer covering construction. Enlarging the balls in the outer construction by a fixed factor makes its levels nested. The two intersections of nested finite unions satisfy

K‾⊆⋂j≥kFc(Tj)⊆K‾.\overline{K} \subseteq\bigcap_{j \ge k} F_c(T_j) \subseteq\overline{K}.

In both constructions, the number of chosen children of each level-jj parent is comparable to

Rj(c):=NA(cψ(Tj),Tj+1)NA(cψ(Tj+1),Tj+1).\mathcal{R}_j(c) := \frac{N_A(c\psi(T_j),T_{j+1})}{N_A(c\psi(T_{j+1}),T_{j+1})}.

Their radii at level NN are comparable to cψ(TN)c\psi(T_N). Let C≥1C \ge1 be a constant, independent of the level jj and of sufficiently large LL, that bounds both the ratio of the number of chosen children to Rj(c)\mathcal{R}_j(c) and its reciprocal.

  • Upper bound. The level-NN cover of K‾\overline{K} has at most a constant multiple of CN−k∏j=kN−1Rj(c)C^{N-k}\prod_{j=k}^{N-1}\mathcal{R}_j(c) balls. Its ss-dimensional covering cost is therefore at most a constant multiple of

CN−k∏j=kN−1Rj(c)⏟Number of balls in the step-N covering(cψ(TN))s.\underbrace{C^{N-k}\prod_{j=k}^{N-1}\mathcal{R}_j(c)}_{\text{Number of balls in the step-}N\text{ covering}}(c\psi(T_N))^s.

Taking logarithms and using −log⁡ψ(TN)=NL-\log\psi(T_N)=NL shows that this tends to zero along a subsequence when s>PA,ψ(L,c)+log⁡C/Ls>\mathcal{P}_{A,\psi}(L,c)+\log C/L. Hence dim⁡HK‾≤PA,ψ(L,c)+O(L−1)\dim_H\overline{K}\le\mathcal{P}_{A,\psi}(L,c)+O(L^{-1}).

  • Lower bound. We use the mass distribution principle to obtain a lower bound for dim⁡HK\dim_H K. Split each parent’s mass equally among its children. Each level-NN ball has mass at most CN−k/∏j=kN−1Rj(c)C^{N-k}/\prod_{j=k}^{N-1}\mathcal{R}_j(c). Disjointness bounds how many such balls meet a test ball of comparable radius. The mass distribution estimate for this construction gives dim⁡HK≥PA,ψ(L,c)−O(L−1)\dim_H K\ge\mathcal{P}_{A,\psi}(L,c)-O(L^{-1}).

In both bounds, fixed comparison factors contribute O(N)O(N) to the logarithm, hence O(L−1)O(L^{-1}) after division by NLNL.

These estimates concern the discrete intersection ⋂j≥kFc(Tj)\bigcap_{j\ge k}F_c(T_j). To pass to all sufficiently large TT, observe that

DA(cψ)=⋃k⋂T≥TkFc(T)⊆⋃k⋂j≥kFc(Tj).D_A(c\psi)=\bigcup_k\bigcap_{T\ge T_k}F_c(T)\subseteq\bigcup_k\bigcap_{j\ge k}F_c(T_j).

Countable stability gives dim⁡HDA(cψ)≤PA,ψ(L,c)+O(L−1)\dim_H D_A(c\psi)\le\mathcal{P}_{A,\psi}(L,c)+O(L^{-1}). For the lower bound, if Tj≤T<Tj+1T_j\le T<T_{j+1}, the denominator chosen at TjT_j is still at most TT, and its error is at most

cψ(Tj)≤ρcψ(Tj+1)≤ρcψ(T),ρ=eL.c\psi(T_j)\le\rho c\psi(T_{j+1})\le\rho c\psi(T),\qquad\rho=e^L.

Thus

K⊆⋂j≥kFc(Tj)⊆DA(ρcψ),K \subseteq\bigcap_{j\ge k} F_c(T_j) \subseteq D_A(\rho c\psi),

giving dim⁡HDA(ρcψ)≥PA,ψ(L,c)−O(L−1)\dim_H D_A(\rho c\psi) \ge P_{A,\psi}(L,c)-O(L^{-1}).

The upper and lower bounds apply to DA(cψ)D_A(c\psi) and DA(ρcψ)D_A(\rho c\psi), respectively. The discrepancy in constants disappears for Sing⁡A(ψ)\operatorname{Sing}_A(\psi) and UA(ψ)U_A(\psi), since their definitions take intersections and unions over all positive constants. Countable stability gives the dimension of UA(ψ)U_A(\psi). For Sing⁡A(ψ)\operatorname{Sing}_A(\psi), this gives only the upper bound dim⁡HSing⁡A(ψ)≤P−(A,ψ)\dim_H \operatorname{Sing}_A(\psi) \le P_-(A,\psi), since an intersection can have smaller dimension. A separate no-dimension-drop argument [16 Proposition 5.4] proves equality, yielding the formula for Sing⁡A(ψ)\operatorname{Sing}_A(\psi) in Theorem 2.3.

Applications. For asymptotic approximation, the Hausdorff-measure tests in § 2.4 follow from the same geometric considerations. The convergence test uses the ss-dimensional covering cost of each good cell; the divergence test enlarges the targets and applies mass transference [2].

For uniform approximation, the successive-minima expression (5) yields the rank formula (11) of Theorem 2.4 for powers outside Em,nE_{m,n}. The critical power ψ(T)=T−n/m\psi(T)=T^{-n/m}, for which the approximation boxes have constant volume (ψ(T)mTn=1)(\psi(T)^mT^n=1), is treated separately: an upper estimate using the small successive minima gives the escape-of-mass bound in Theorem 2.5. DEPARTMENT OF MATHEMATICS, BRANDEIS UNIVERSITY, WALTHAM, MA 02453, USA

Email address: [email protected]

DEPARTMENT OF MATHEMATICS, BRANDEIS UNIVERSITY, WALTHAM, MA 02453, USA

Email address: [email protected]

References

  1. [1]Gaurav Aggarwal, Dimension bounds for singular affine forms, Mathematika 72 (2026), e70084.DOI
  2. [2]Victor Beresnevich and Sanju Velani, A mass transference principle and the Duffin–Schaeffer conjecture for Hausdorff measures, Annals of Mathematics 164 (2006), no. 3, 971–992.arxiv.org/abs/math/0412141
  3. [3]Victor Beresnevich and Sanju Velani, Classical metric Diophantine approximation revisited: The Khintchine–Groshev theorem, International Mathematics Research Notices 2010 (2010), no. 1, 69–86.arxiv.org/abs/0811.0809
  4. [4]Yann Bugeaud and Michel Laurent, On exponents of homogeneous and inhomogeneous Diophantine approximation, Moscow Mathematical Journal 5 (2005), no. 4, 747–766.DOI
  5. [5]Tushar Das, Lior Fishman, David Simmons, and Mariusz Urbański, A variational principle in the parametric geometry of numbers, Advances in Mathematics 437 (2024), 109435.DOI
  6. [6]Michael Fuchs and Dong Han Kim, On Kurzweil’s 0–1 law in inhomogeneous Diophantine approximation, Acta Arithmetica 173 (2016), no. 1, 41–57.arxiv.org/abs/1501.04714
  7. [7]A. V. Groshev, A theorem on a system of linear forms, Doklady Akademii Nauk SSSR 19 (1938), no. 3, 151–152 (in Russian).
  8. [8]Martin Henk, Successive minima and lattice points, Rendiconti del Circolo Matematico di Palermo. Serie II. Supplemento (2002), no. 70, part I, 377–384.DOI
  9. [9]Mumtaz Hussain and Benjamin Ward, Weighted twisted inhomogeneous Diophantine approximation, Nonlinearity 37 (2024), 125004.arxiv.org/abs/2307.13210
  10. [11]A. Khintchine, Zur metrischen Theorie der diophantischen Approximationen, Mathematische Zeitschrift 24 (1926), 706–714.DOI
  11. [12]Taehyeong Kim, On a Kurzweil type theorem via ubiquity, Acta Arithmetica 213 (2024), no. 2, 181–191.arxiv.org/abs/2306.00847
  12. [13]Dong Han Kim and Lingmin Liao, Dirichlet uniformly well-approximated numbers, International Mathematics Research Notices 2019 (2019), no. 24, 7691–7732.arxiv.org/abs/1508.00520
  13. [14]Taehyeong Kim and Wooyeon Kim, Hausdorff measure of sets of Dirichlet non-improvable affine forms, Advances in Mathematics 403 (2022), 108353.arxiv.org/abs/2006.05727
  14. [15]Taehyeong Kim and Vasiliy Neckrasov, Twisted Diophantine approximation I: Asymptotic theory, arXiv:2609.36411, 2026.arxiv.org/abs/2609.36411
  15. [16]———, Twisted Diophantine approximation II: Uniform theory, arXiv:2609.36412, 2026.arxiv.org/abs/2609.36412
  16. [17]Dong Han Kim, Michal Rams, and Baowei Wang, Hausdorff dimension of the set approximated by irrational rotations, Mathematika 64 (2018), no. 1, 267–283.DOI
  17. [18]Dmitry Kleinbock, Andreas Strömbergsson, and Shucheng Yu, A measure estimate in geometry of numbers and improvements to Dirichlet’s theorem, Proceedings of the London Mathematical Society (3) 125 (2022), 778–824.arxiv.org/abs/2108.04638
  18. [19]Dmitry Kleinbock and Nick Wadleigh, A zero-one law for improvements to Dirichlet’s theorem, Proceedings of the American Mathematical Society 146 (2018), no. 5, 1833–1844.arxiv.org/abs/1609.06780
  19. [20]Dmitry Kleinbock and Nick Wadleigh, An inhomogeneous Dirichlet theorem via shrinking targets, Compositio Mathematica 155 (2019), 1402–1423.DOI
  20. [21]Jaroslav Kurzweil, On the metric theory of inhomogeneous Diophantine approximations, Studia Mathematica 15 (1955), no. 1, 84–112.DOI
  21. [22]Nikolay Moshchevitin and Vasiliy Neckrasov, Metric theory of inhomogeneous Diophantine approximations with a fixed matrix, arXiv:2503.21180, 2025.arxiv.org/abs/2503.21180
  22. [25]Wolfgang M. Schmidt and Leonhard Summerer, Parametric geometry of numbers and applications, Acta Arithmetica 140 (2009), no. 1, 67–91.DOI
  23. [26]Vladimir G. Sprindžuk, Metric theory of Diophantine approximations, John Wiley & Sons, New York, 1979. English translation.
  24. [27]Andreas Strömbergsson and Shucheng Yu, A zero-one law for improvements to Dirichlet’s theorem in arbitrary dimension, arXiv:2602.16258, 2026.arxiv.org/abs/2602.16258

Paper details

vlnc:W_axk94f2a
Version
v1 · latest
Published
Fresh — added to the Hub within the last 30 days
Added
License
CC BY 4.0
Usage

views · 0

downloads · 0

A view counts once per browser per day, a download each time the PDF is requested.

  • Mathematics

Not recorded for this paper

View statement
Looking for related papers…
Contents