Hostname: page-component-78c5997874-fbnjt Total loading time: 0 Render date: 2024-11-08T13:24:41.901Z Has data issue: false hasContentIssue false

ON THE TWO-PARAMETER ERDŐS–FALCONER DISTANCE PROBLEM IN FINITE FIELDS

Published online by Cambridge University Press:  29 September 2022

FRANCOIS CLÉMENT
Affiliation:
Sorbonne Université, CNRS, LIP6, Paris, France e-mail: [email protected]
HOSSEIN NASSAJIAN MOJARRAD
Affiliation:
Courant Institute, New York University, New York 10012, USA e-mail: [email protected]
DUC HIEP PHAM*
Affiliation:
University of Education, Vietnam National University, Hanoi, Vietnam
CHUN-YEN SHEN
Affiliation:
Department of Mathematics, National Taiwan University, Taipei 10617, Taiwan e-mail: [email protected]
*
Rights & Permissions [Opens in a new window]

Abstract

Given $E \subseteq \mathbb {F}_q^d \times \mathbb {F}_q^d$ , with the finite field $\mathbb {F}_q$ of order q and the integer $d\,\ge \, 2$ , we define the two-parameter distance set $\Delta _{d, d}(E)=\{(\|x-y\|, \|z-t\|) : (x, z), (y, t) \in E \}$ . Birklbauer and Iosevich [‘A two-parameter finite field Erdős–Falconer distance problem’, Bull. Hellenic Math. Soc. 61 (2017), 21–30] proved that if $|E| \gg q^{{(3d+1)}/{2}}$ , then $ |\Delta _{d, d}(E)| = q^2$ . For $d=2$ , they showed that if $|E| \gg q^{{10}/{3}}$ , then $ |\Delta _{2, 2}(E)| \gg q^2$ . In this paper, we give extensions and improvements of these results. Given the diagonal polynomial $P(x)=\sum _{i=1}^da_ix_i^s\in \mathbb F_q[x_1,\ldots , x_d]$ , the distance induced by P over $\mathbb {F}_q^d$ is $\|x-y\|_s:=P(x-y)$ , with the corresponding distance set $\Delta ^s_{d, d}(E)=\{(\|x-y\|_s, \|z-t\|_s) : (x, z), (y, t) \in E \}$ . We show that if $|E| \gg q^{{(3d+1)}/{2}}$ , then $ |\Delta _{d, d}^s(E)| \gg q^2$ . For $d=2$ and the Euclidean distance, we improve the former result over prime fields by showing that $ |\Delta _{2,2}(E)| \gg p^2$ for $|E| \gg p^{{13}/{4}}$ .

Type
Research Article
Copyright
© The Author(s), 2022. Published by Cambridge University Press on behalf of Australian Mathematical Publishing Association Inc.

1 Introduction

The general Erdős distance problem is to determine the number of distinct distances spanned by a finite set of points. In the Euclidean space, it is conjectured that for any finite set $E \subset \mathbb {R}^d$ , $d\ge 2$ , we have $|\Delta (E)| \gtrapprox |E|^{{2}/{d}}$ , where $\Delta (E)=\{\|x-y\| : x,y \in E\}$ . Here and throughout, $X\ll Y$ means that there exists $C>0$ such that $X\le CY$ , and $X\lessapprox Y$ with the parameter N means that for any $\varepsilon>0$ , there exists $C_{\varepsilon }>0$ such that $X\,\le \,C_{\varepsilon }N^{\varepsilon }Y$ .

The finite field analogue of the distance problem was first studied by Bourgain et al. [Reference Bourgain, Katz and Tao2] over prime fields. In this setting, the Euclidean distance between any two points $x=(x_1,\ldots , x_d), y=(y_1,\ldots ,y_d) \in \mathbb {F}_q^d$ , the d-dimensional vector space over the finite field $\mathbb F_q$ of order q, is $\|x-y\|=\sum _{i=1}^d(x_i-y_i)^2\in \mathbb F_q$ . For prime fields $\mathbb {F}_p$ with $p\equiv 1 \pmod 4$ , they showed that if $E \subset \mathbb {F}_p^2$ with $|E|=p^{\delta }$ for some $0<\delta <2$ , then the distance set satisfies $|\Delta (E)| \gg |E|^{{1}/{2}+\varepsilon }$ for some $\varepsilon>0$ depending only on $\delta $ .

This bound does not hold in general for arbitrary finite fields $\mathbb {F}_q$ , as shown by Iosevich and Rudnev [Reference Iosevich and Rudnev9]. In this general setting, they considered the Erdős–Falconer distance problem to determine how large $E \subset \mathbb {F}_q^d$ needs to be so that $\Delta (E)$ spans all possible distances or at least a positive proportion of them. More precisely, they proved that $\Delta (E)=\mathbb {F}_q$ if $|E|> 2q^{{(d+1)}/{2}}$ in the all distances case, and also conjectured that $|\Delta (E)| \gg q$ if $|E| \gg _{\varepsilon } q^{{d}/{2}+\varepsilon }$ in the positive proportion case. In [Reference Hart, Iosevich, Koh and Rudnev6], it was shown that the exponent in the all distances case is sharp for odd d, and the conjecture for the positive proportion case holds for all $E \subset \{x\in \mathbb {F}_q^d : \|x\|=1\}$ . It is conjectured that in even dimensions, the optimal exponent is ${d}/{2}$ for the all distances case. In particular for $d=2$ , it was shown in [Reference Chapman, Erdogan, Hart, Iosevich and Koh3] that if $E \subseteq \mathbb {F}_q^2$ satisfies $|E| \gg q^{{4}/{3}}$ , then $|\Delta (E)| \gg q$ , improving the positive proportion case. The proofs in [Reference Chapman, Erdogan, Hart, Iosevich and Koh3] use extension estimates for circles. Therefore, one would expect to get improvements for distance problems if one can obtain improved estimates for extension problems.

There have been a recent series of other improvements and generalisations on the Erdős–Falconer distance problem. In [Reference Hieu and Pham7], a generalisation for subsets of regular varieties was studied. Extension theorems and their connection to the Erdős–Falconer problem are the main focus of [Reference Koh, Pham and Vinh10]. The exponents ${(d+1)}/{2}$ and ${d}/{2}$ were improved in [Reference Pham and Suk13, Reference Pham and Vinh14] for subsets E with Cartesian product structure in the all distances case for $|\Delta (E)|$ and in the positive proportion case for the quotient distance set $|{\Delta (E)}/{\Delta (E)}|$ .

A two-parameter variant of the Erdős–Falconer distance problem for the Euclidean distance was studied by Birklbauer and Iosevich in [Reference Birklbauer and Iosevich1]. More precisely, given $E \subseteq \mathbb {F}_q^d \times \mathbb {F}_q^d$ , where $d\ge 2$ , define the two-parameter distance set as

$$ \begin{align*} \Delta_{d, d}(E)=\{(\|x-y\|, \|z-t\|) : (x,z), (y,t) \in E \} \subseteq \mathbb{F}_q \times \mathbb{F}_q. \end{align*} $$

They proved the following results.

Theorem 1.1 [Reference Birklbauer and Iosevich1]

Let E be a subset in $\mathbb {F}_q^d \times \mathbb {F}_q^d$ . If $|E| \gg q^{{(3d+1)}/{2}}$ , then $ |\Delta _{d, d}(E)| = q^2$ .

Theorem 1.2 [Reference Birklbauer and Iosevich1]

Let E be a subset in $\mathbb {F}_q^2 \times \mathbb {F}_q^2$ . If $|E| \gg q^{{10}/{3}}$ , then $ |\Delta _{2, 2}(E)| \gg q^2$ .

In this short note, we provide an extension and an improvement of these results. Unlike [Reference Birklbauer and Iosevich1], which relies heavily on Fourier analytic techniques, we use an elementary counting approach.

Let $P(x)=\sum _{i=1}^da_ix_i^s\in \mathbb F_q[x_1,\ldots , x_d]$ be a fixed diagonal polynomial in d variables of degree $s\ge 2$ . For $x=(x_1,\ldots ,x_d), y=(y_1,\ldots ,y_d) \in \mathbb {F}_q^d$ , we introduce

$$ \begin{align*} \|x-y\|_s:=P(x-y)=\sum_{i=1}^d a_i(x_i-y_i)^s\in\mathbb F_q. \end{align*} $$

For any set $E\subset \mathbb {F}_q^d\times \mathbb {F}_q^d$ , define

$$ \begin{align*} \Delta_{d, d}^s(E)=\{(\|x-y\|_s, \|z-t\|_s) : (x, z), (y, t) \in E \}\in\mathbb F_q\times\mathbb F_q. \end{align*} $$

Our first result reads as follows.

Theorem 1.3. Let E be a subset in $\mathbb {F}_q^d \times \mathbb {F}_q^d$ . If $|E| \gg q^{{(3d+1)}/{2}}$ , then $ |\Delta _{d, d}^s(E)| \gg q^2$ .

Our method also works for the multi-parameter distance set for $E \subseteq \mathbb {F}_q^{d_1+\cdots +d_k}$ , but we do not discuss such extensions here. For $d=2$ , we get an improved version of Theorem 1.2 for the Euclidean distance function over prime fields.

Theorem 1.4. Let $E \subseteq \mathbb {F}_p^2 \times \mathbb {F}_p^2$ . If $|E| \gg p^{{13}{/4}}$ , then $ |\Delta _{2,2}(E)| \gg p^2$ .

The continuous versions of Theorems 1.3 and 1.4 have been studied in [Reference Du, Ou and Zhang4, Reference Hambrook, Iosevich and Rice5, Reference Iosevich, Janczak and Passant8]. We do not know whether our method can be extended to that setting. It follows from our approach that the conjectured exponent ${d}/{2}$ of the (one-parameter) distance problem would imply the sharp exponent for the two-parameter analogue, namely ${3d}/{2}$ , for even dimensions. We refer to [Reference Birklbauer and Iosevich1] for constructions and more discussions.

2 Proof of Theorem 1.3

By using the following auxiliary result whose proof relies on Fourier analytic methods (see [Reference Vinh15, Theorem 2.3] and [Reference Koh and Shen11, Corollaries 3.1 and 3.4]), we are able to give an elegant proof for Theorem 1.3. Compared with the method in [Reference Birklbauer and Iosevich1], ours is more elementary.

Lemma 2.1. Let $X, Y \subseteq \mathbb {F}_q^d$ . Define $\Delta ^s(X, Y)=\{\|x-y\|_s\colon x\in X, y\in Y\}$ . If $|X||Y|\gg q^{d+1}$ , then $|\Delta ^s(X, Y)|\gg q$ .

Proof of Theorem 1.3

By assumption, $|E|\ge Cq^{d+{(d+1)}/{2}}$ for some constant $C>0$ . For $y\in \mathbb {F}_q^d$ , let $E_y:=\{x\in \mathbb F_q^d : (x, y)\in E\}$ and define

$$ \begin{align*} Y:=\big\{ y\in \mathbb{F}_q^d : |E_y|> \tfrac{1}{2}C q^{{(d+1)}/{2}} \big\}. \end{align*} $$

We first show that $ |Y|\,\ge \, \tfrac 12C q^{{(d+1)}/{2}}$ . Note that

$$ \begin{align*}|E|=\sum_{y \in Y} |E_y| + \sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y| \,\le\, q^d|Y| + \sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y|, \end{align*} $$

where the last inequality holds since $|E_y|\,\le \,q^d$ for $y\in \mathbb {F}_q^d$ . Combining it with the assumption on $|E|$ gives the lower bound

$$ \begin{align*} \displaystyle\sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y|\,\ge\, Cq^{d+{(d+1)}/{2}} - q^d|Y|. \end{align*} $$

However, by definition, $|E_y|\,\le \,\tfrac 12C q^{{(d+1)}/{2}}$ for $y \in \mathbb {F}^d_q\setminus Y$ , yielding the upper bound

$$ \begin{align*} \displaystyle\sum_{y \in \mathbb{F}^d_q\setminus Y} |E_y|\,\le\,\dfrac{1}{2}Cq^{d+{(d+1)}/{2}} .\end{align*} $$

These two bounds together give $Cq^{d+{(d+1)}/{2}} - q^d|Y|\,\le \,\tfrac 12C q^{d+{(d+1)}/{2}}$ , proving the claimed bound $|Y|\,\ge \, \tfrac 12C q^{{(d+1)}/{2}}$ .

In particular, Lemma 2.1 implies $|\Delta ^s(Y,Y)|\gg q$ , as $|Y||Y| \gg q^{d+1}$ . However, for each $u\in \Delta ^s(Y,Y)$ , there are $z, t\in Y$ such that $\|z-t\|_s=u$ . One has $|E_z|, |E_t|\gg q^{{(d+1)}/{2}}$ , therefore, again by Lemma 2.1, $|\Delta ^s(E_z, E_t)|\gg q$ . Furthermore, for $v\in \Delta ^s(E_z, E_t)$ , there are $x\in E_z$ and $y\in E_t$ satisfying $\|x-y\|_s=v$ . Note that $x\in E_z$ and $y\in E_t$ mean that $(x,z), (y,t)\in E$ . Thus, $(v,u)=(\|x-y\|_s, \|z-t\|_s)\in \Delta _{d, d}^s(E)$ . From this, we conclude that $|\Delta _{d, d}^s(E)|\gg q|\Delta ^s(Y,Y)|\gg q^2$ , which completes the proof.

3 Proof of Theorem 1.4

To improve the exponent over prime fields $\mathbb {F}_p$ , we strengthen Lemma 2.1 as shown in Lemma 3.1 below. Following the proof of Theorem 1.3 and using Lemma 3.1 proves Theorem 1.4.

Lemma 3.1. Let $X, Y \subseteq \mathbb {F}_p^2$ . If $|X|, |Y|\gg p^{5/4}$ , then $|\Delta (X, Y)|\gg p$ .

Proof. It is clear that if $X' \subseteq X$ and $Y' \subseteq Y$ , then $\Delta (X',Y')\subseteq \Delta (X,Y)$ . Thus, without loss of generality, we may assume that $|X|=|Y|=N$ with $N\gg p^{5/ 4}$ . Let Q be the number of quadruples $(x, y, x', y')\in X\times Y\times X\times Y$ such that $\|x-y\|=\|x'-y'\|$ . It follows easily from the Cauchy–Schwarz inequality that

$$ \begin{align*}|\Delta(X, Y)|\gg \frac{|X|^2|Y|^2}{Q}.\end{align*} $$

Let T be the number of triples $(x, y, y')\in X\times Y\times Y $ such that $\|x-y\|=\|x-y'\|$ . By the Cauchy–Schwarz inequality again, one gets $Q\ll |X|\cdot T$ . Next, we need to bound T. For this, denote $Z=X\cup Y$ , so that $N\le |Z|\le 2N$ . Let $T'$ be the number of triples $(a, b, c)\in Z\times Z\times Z$ such that $\|a-b\|=\|a-c\|$ . Obviously, $T\le T'$ . However, it was recently proved (see [Reference Murphy, Petridis, Pham, Rudnev and Stevens12, Theorem 4]) that

$$ \begin{align*}T'\ll \frac{|Z|^3}{p}+p^{2 /3}|Z|^{ 5 /3}+p^{1 /4}|Z|^2,\end{align*} $$

which gives

$$ \begin{align*}T\ll \frac{N^3}{p}+p^{2 /3}N^{5 /3}+p^{1 /4}N^2,\end{align*} $$

and then $T\ll {N^3}/p$ (since $N\gg p^{5 /4}$ ). Putting all the bounds together, we obtain

$$ \begin{align*}\dfrac{N^3}{|\Delta(X,Y)|}=\dfrac{|X||Y|^2}{|\Delta(X,Y)|}\ll \dfrac Q{|X|}\ll T\ll \dfrac {N^3}p,\end{align*} $$

or equivalently, $|\Delta (X,Y)|\gg p$ , as required.

Acknowledgement

The authors are grateful to Dr. Thang Pham for sharing insights and new ideas.

Footnotes

The second-listed author was supported by Swiss National Science Foundation grant P2ELP2-178313.

References

Birklbauer, P. and Iosevich, A., ‘A two-parameter finite field Erdős–Falconer distance problem’, Bull. Hellenic Math. Soc. 61 (2017), 2130.Google Scholar
Bourgain, J., Katz, N. and Tao, T., ‘A sum-product estimate in finite fields, and applications’, Geom. Funct. Anal. 14 (2004), 2757.CrossRefGoogle Scholar
Chapman, J., Erdogan, M., Hart, D., Iosevich, A. and Koh, D., ‘Pinned distance sets, $k$ -simplices, Wolff’s exponent in finite fields and sum-product estimates’, Math. Z. 271 (2012), 6393.CrossRefGoogle Scholar
Du, X., Ou, Y. and Zhang, R., ‘On the multiparameter Falconer distance problem’, Trans. Amer. Math. Soc. 375 (2022), 49795010.CrossRefGoogle Scholar
Hambrook, K., Iosevich, A. and Rice, A., ‘Group actions and a multi-parameter Falconer distance problem’, Preprint, 2017, arXiv:1705.03871.Google Scholar
Hart, D., Iosevich, A., Koh, D. and Rudnev, M., ‘Averages over hyperplanes, sum-product theory in vector spaces over finite fields and the Erdös–Falconer distance conjecture’, Trans. Amer. Math. Soc. 363 (2011), 32553275.CrossRefGoogle Scholar
Hieu, D. and Pham, T., ‘Distinct distances on regular varieties over finite fields’, J. Number Theory 173 (2017), 602613.CrossRefGoogle Scholar
Iosevich, A., Janczak, M. and Passant, J., ‘A multi-parameter variant of the Erdős distance problem’, Preprint, 2017, arXiv:1712.04060.Google Scholar
Iosevich, A. and Rudnev, M., ‘Erdős distance problem in vector spaces over finite fields’, Trans. Amer. Math. Soc. 359 (2007), 61276142.CrossRefGoogle Scholar
Koh, D., Pham, T. and Vinh, L., ‘Extension theorems and a connection to the Erdős–Falconer distance problem over finite fields’, J. Funct. Anal. 281 (2021), 154.CrossRefGoogle Scholar
Koh, D. and Shen, C.-Y., ‘The generalized Erdős–Falconer distance problems in vector spaces over finite fields’, J. Number Theory 132 (2012), 24552473.CrossRefGoogle Scholar
Murphy, B., Petridis, G., Pham, T., Rudnev, M. and Stevens, S., ‘On the pinned distances problem in positive characteristic’, J. Lond. Math. Soc. (2) 105 (2022), 469499.Google Scholar
Pham, T. and Suk, A., ‘Structures of distance sets over prime fields’, Proc. Amer. Math. Soc. 148 (2020), 32093215.CrossRefGoogle Scholar
Pham, T. and Vinh, L., ‘Distribution of distances in vector spaces over prime fields’, Pacific J. Math. 309 (2020), 437451.CrossRefGoogle Scholar
Vinh, L., ‘On the generalized Erdős–Falconer distance problems over finite fields’, J. Number Theory 133 (2013), 29392947.CrossRefGoogle Scholar