Improvements on exponential sums related to Piatetski-Shapiro primes

Li Lu, Lingyu Guo, Victor Zhenyu Guo · 2025

Read Improvements on exponential sums related to Piatetski-Shapiro primes by Li Lu, Lingyu Guo, Victor Zhenyu Guo (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-2504-11464.txt.

Introduction

Li Lu, Lingyu Guo, and Victor Guo sharpen bounds on a kind of weighted sum used in the study of special primes, in a paper from 2025. The primes in question are the whole parts of powers of whole numbers. Such primes, named for Piatetski and Shapiro, follow a sparse pattern. Sums of this kind are the main tool for proving where they fall.

Their new bound comes from a fresh attack on one part of the sum. With it, the range of a key condition from Balog and Friedlander grows wider. That wider range improves a known result on writing numbers as sums of three such primes. The methods also bear on how these primes spread across arithmetic patterns. They even touch primes shared by several beat-like sequences at once. The paper is a step in a long analytic program.

Catalog record

CREATORLi Lu, Lingyu Guo, Victor Zhenyu Guo
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE37 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2504-11464

Browse the full library · About the Tower of Babel

Full text

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

Improvements on exponential sums related to Piatetski-Shapiro primes Li Lu, Lingyu Guo, Victor Zhenyu Guo Abstract . We prove a new bound to the exponential sum of the form ∑ h ∼ H δ h ​ ∑ m ∼ M ∑ n ∼ N m ​ n ∼ x ⁡ a m ​ b n ​ 𝐞 ​ ( α ​ m ​ n + h ​ ( m ​ n + u ) γ ) , \sum_{h\sim H}\delta_{h}\mathop{\sum_{m\sim M}\sum_{n\sim N}}_{mn\sim x}a_{m}b_{n}\mathbf{e}\big(\alpha mn+h(mn+u)^{\gamma}\big), by a new approach to the Type I sum. The sum can be applied to many problems related to Piatetski-Shapiro primes, which are primes of the form ⌊ n c ⌋ \lfloor n^{c}\rfloor . In this paper, we improve the admissible range of the Balog-Friedlander condition, which leads to an improvement to the ternary Goldbach problem with Piatetski-Shapiro primes. We also investigate the distribution of Piatetski-Shapiro primes in arithmetic progressions, Piatetski-Shapiro primes in the intersection of multiple Beatty sequences and so on. MSC Numbers: 11N05; 11B83; 11L07. Keywords: exponential sums; Piateski-Shapiro primes; ternary Goldbach problem. 1. Introduction and main results The Piatetski-Shapiro sequences are sequences of the form \mathcal{N}^{(c)}\mathrel{\vbox{\hbox{\scriptsize.}\hbox{\scriptsize.}}}=(\lfloor n^{c}\rfloor)_{n=1}^{\infty}, where \left\lfloor\cdot\right\rfloor is the integer part. Piatetski-Shapiro [ 12 ] proved the Piatetski-Shapiro prime number theorem stating that for 1<c<\frac{12}{11} the counting function \pi^{(c)}(x)\mathrel{\vbox{\hbox{\scriptsize.}\hbox{\scriptsize.}}}=\#\big\{\text{\rm prime~}p\leqslant x:p\in\mathcal{N}^{(c)}\big\} satisfies the asymptotic relation (1.1) \pi^{(c)}(x)=(1+o(1))\frac{x^{1/c}}{\log x}\qquad\text{~\rm as }x\to\infty. The admissible range for c of the above formula has been extended many times and is currently known to hold for all 1<c<\frac{2817}{2426} thanks to Rivat and Sargos [ 14 ] . Rivat and Wu [ 15 ] also showed that there are infinitely many Piatetski-Shapiro primes for 1<c<\frac{243}{205} without an asymptotic formula. We refer the readers to [ 7 ] for more details of the improvements of c . The asymptotic relation is expected to hold for all values of 1<c<2 . The estimation of Piatetski-Shapiro primes is an approximation of the well-known conjecture that there exist infinitely many primes of the form n^{2}+1 . In this article, we mainly consider three types of problems related to Piatetski-Shapiro primes by proving a new bound on an special type of exponential sums which improves the bound by Kumchev [ 11 ] . Several results are improved based on our new estimation of the exponential sum. Throughout the paper, we always donote \gamma\mathrel{\vbox{\hbox{\scriptsize.}\hbox{\scriptsize.}}}=c^{-1} for any subscripts. 1.1. The ternary Goldbach problem on Piatetski-Shapiro primes A natural problem to study prime numbers is the ternary Goldbach problem, which is proved by Vinogradov that every odd sufficiently large integer is a sum of three primes. In 1992, Balog and Friedlander [ 2 ] considered the ternary Goldbach problem on Piatetski-Shapiro primes and proved that for every sufficiently large N (1.2) \sum_{\begin{subarray}{c}p_{1}+p_{2}+p_{3}=N\\ p_{i}\in\mathcal{N}^{(c_{i})}\end{subarray}}1=(1+o(1))\frac{\gamma_{1}\gamma_{2}\gamma_{3}\Gamma(\gamma_{1})\Gamma(\gamma_{2})\Gamma(\gamma_{3})}{\Gamma(\gamma_{1}+\gamma_{2}+\gamma_{3})}\frac{\mathfrak{S}(N)N^{\gamma_{1}+\gamma_{2}+\gamma_{3}-1}}{\log^{3}N}, where \Gamma(\cdot) is the Gamma function and \mathfrak{S}(N) is the singular series which is defined as (1.3) \mathfrak{S}(N)\mathrel{\vbox{\hbox{\scriptsize.}\hbox{\scriptsize.}}}=\prod_{p|N}\bigg(1-\frac{1}{(p-1)^{2}}\bigg)\prod_{p\nmid N}\bigg(1+\frac{1}{(p-1)^{3}}\bigg), with 1<c_{1},c_{2},c_{3}<\frac{21}{20} . The admissible range of c was improved several times as the following table. — Authors The upper bound of c Balog and Friedlander [ 2 ] \frac{21}{20}=1.05 Rivat [ 13 ] \frac{199}{188}=1.0585... Kumchev [ 11 ] \frac{53}{50}=1.06 Sun, Pan and Du [ 16 ] \frac{73}{64}=1.1406... We improve the admissible range of c and prove the following theorem. Theorem 1.1 . With 1<c_{1},c_{2},c_{3}<\frac{569}{498}=1.1425\dots , ( 1.2 ) holds. To prove Theorem 1.1 , the key part is the Balog-Friedlander condition due to Sun, Du and Pan. Definition 1.2 (Balog-Friedlander condition) . A real number c\in(1,2) satisfies the Balog-Friedlander condition if for every sufficiently large N , it follows that (1.4) \sum_{\begin{subarray}{c}p\leqslant N\\ p\in\mathcal{N}^{(c)}\end{subarray}}cp^{1-\frac{1}{c}}\log p\cdot\mathbf{e}(\alpha p)=\sum_{p\leqslant N}\log p\cdot\mathbf{e}(\alpha p)+O_{\varepsilon}(N^{1-\varepsilon}), uniformly for \alpha . ∎ Balog and Friedlander proved that the condition holds for 1<c<\frac{9}{8}=1.125 while Kumchev improved it to 1<c<\frac{73}{64}=1.1406... . We obtain the following theorem. Theorem 1.3 . For 1<c<\frac{569}{498}=1.1425\dots , the Balog-Friedlander condition holds. Remark. We mention that there is a unsolvable mistake in the proof of the Balog-Friedlander condition by Kumchev [ 11 ] , which causes that the result might be wrong. Since the condition is the key part to prove ( 1.2 ), the result by Kumchev does not follow. We give a detailed explanation of the mistake in Section 3 . 1.2. Piatetski-Shapiro primes in arithmetic progressions Another natural topic related to primes is the study of primes in arithmetic progressions. For any integers a and q with (a,q)=1 , we define the counting function \pi^{(c)}(x;q,a):=\#\big\{p\leqslant x:p\equiv a\pmod{q},\,p\in\mathcal{N}^{(c)}\big\} in terms of the more familiar function \pi(x;q,a)\mathrel{\vbox{\hbox{\scriptsize.}\hbox{\scriptsize.}}}=\#\big\{p\leqslant x:p\equiv a\pmod{q}\big\}. Baker, Banks, Brüdern, Shparlinski and Weingartner [ 3 ] proved an asymptotic formula and a bound of error term to \pi^{(c)}(x;q,a) . We list all the investigations of the admissble range of c in the following table. — Authors The range of c Baker et al. [ 3 ] 1<c<\frac{18}{17}=1.0588\dots Guo [ 6 ] 1<c<\frac{14}{13}=1.0769\dots Guo, Li and Zhang [ 9 ] 1<c<\frac{12}{11}=1.0909\dots In this paper, we continue improving the range of c and establish the following result. Theorem 1.4 . Let a and q be coprime integers. For 1<c<\frac{569}{498}=1.1425\dots , we have \pi^{(c)}(x;q,a)=\gamma x^{\gamma-1}\pi(x;q,a)+\gamma(1-\gamma)\intop\nolimits_{2}^{x}u^{\gamma-2}\pi(u;q,a)\mathrm{d}u+O(x^{\gamma-\varepsilon}), where the implied constant depends on \gamma and \varepsilon . 1.3. Piatetski-Shapiro primes in the intersection of multiple Beatty sequences For fixed real numbers \alpha and \beta , the associated non–homogeneous Beatty sequence is the sequence of integers defined by \mathcal{B}_{\alpha,\beta}\mathrel{\vbox{\hbox{\scriptsize.}\hbox{\scriptsize.}}}=\left(\left\lfloor\alpha n+\beta\right\rfloor\right)_{n=1}^{\infty}, which are also called generalized arithmetic progressions. Guo, Li and Zhang [ 8 ] proved that there are infinitely many Piatetski-Shapiro primes in the intersection of a finite number of Beatty sequences with some certain restrictions for 1<c<\frac{12}{11} . With a new estimation of the related exponential sums, we extend the admissible range of \gamma . Theorem 1.5 . Suppose that \xi is a positive integer, and \alpha_{1},\dots,\alpha_{\xi},\beta_{1},\dots,\beta_{\xi}\in\mathbb{R} . Let \alpha_{1},\dots,\alpha_{\xi}>1 be irrational and of finite type such that 1,\alpha_{1}^{-1},\dots,\alpha_{\xi}^{-1}\text{ are linearly independent over }\mathbb{Q}. For 1<c<\frac{569}{498}=1.1425\dots , there are infinitely many primes in the intersection of Beatty sequences \mathcal{B}_{\alpha_{1},\beta_{1}},\dots,\mathcal{B}_{\alpha_{\xi},\beta_{\xi}} and the Piatetski–Shapiro sequence \mathcal{N}^{(c)} . Moreover, the counting function \pi^{(c)}_{\alpha_{1},\beta_{1};\dots;\alpha_{\xi},\beta_{\xi}}(x)\mathrel{\vbox{\hbox{\scriptsize.}\hbox{\scriptsize.}}}=\#\big\{\text{\rm prime~}p\leqslant x:p\in\mathcal{B}_{\alpha_{1},\beta_{1}}\cap\cdots\cap\mathcal{B}_{\alpha_{\xi},\beta_{\xi}}\cap\mathcal{N}^{(c)}\big\} satisfies \pi^{(c)}_{\alpha_{1},\beta_{1};\dots;\alpha_{\xi},\beta_{\xi}}(x)=\frac{x^{\gamma}}{{\alpha_{1}\cdots\alpha_{\xi}}\log x}+O\bigg(\frac{x^{\gamma}}{\log^{2}x}\bigg), where the implied constant depends only on \alpha_{1},\dots,\alpha_{\xi} and \gamma . 1.4. A key result to exponential sums The improvements of our theorems related to Piatetski-Shapiro sequences rely on the following bound. The definition of exponent pair can be found on Page 30 of [ 5 ] . Theorem 1.6 . Let (k,l) be an exponent pair such that 4k-2l+1>0 and \gamma subjected to \displaystyle\frac{12k+10}{12k-2l+13}<\gamma<1. Assume further that H\leqslant x^{1-\gamma+\varepsilon} and 0\leqslant u\leqslant 1 . Then \displaystyle\sum_{h\sim H}\bigg|\sum_{n\sim x}\Lambda(n)\mathbf{e}\big(\alpha n+h(n+u)^{\gamma}\big)\bigg|\ll x^{1-\varepsilon}. Thanks to Bourgain’s powerful exponent pair (\frac{13}{84}+\varepsilon,\frac{55}{84}+\varepsilon) in Theorem 6 of [ 4 ] , we obtain the following corollary. Corollary 1.7 . \displaystyle\sum_{h\sim H}\bigg|\sum_{n\sim x}\Lambda(n)\mathbf{e}\big(\alpha n+h(n+u)^{\gamma}\big)\bigg|\ll x^{1-\varepsilon} provided \frac{498}{569}<\gamma<1 , H\leqslant x^{1-\gamma+\varepsilon} and 0\leqslant u\leqslant 1 . 1.5. Other related topics We should mention that our Corollary 1.7 is very useful while people are investigating combination problems related to Piatetski-Shapiro primes, Beatty primes, primes in arithmetic progressions and the ternary Goldbach problem. Indeed, we believe that following questions can also be proved with the same admissible range: • Piatetski-Shapiro primes in arithmetic progressions in the intersection of Beatty sequences; • The ternary Goldbach problem with primes in the intersection of a Piatetski-Shapiro sequence and Beatty sequences; • The ternary Goldbach problem with Piatetski-Shapiro primes

Read the complete text