On Piatetski-Shapiro sequences related to dense sets

Wei Zhang
DOI: https://doi.org/10.48550/arXiv.2304.10119
2023-04-20
Abstract:In this paper, we consider Piatetski-Shapiro sequences related to dense sets. We can improve a classical result of Rivat and Sárközy. We can enlarge previous effective range $(1,13/11)$ to $(1,6/5).$
Number Theory
What problem does this paper attempt to address?
This paper attempts to solve the problem of dense sets related to the Piatetski - Shapiro sequence and improves the classical results of Rivat and Sárközy. Specifically, the main contribution of the paper is to expand the previous effective range from \((1, \frac{13}{11})\) to \((1, \frac{6}{5})\). ### Problem Background The Piatetski - Shapiro sequence refers to a sequence of the form \([n^c]\), where \(c\) is a constant and \(1 < c < 2\). Piatetski - Shapiro proved that for any \(c\in(1, \frac{12}{11})\), there are infinitely many primes of the form \([n^c]\). This result has been extended by many researchers, and the currently largest acceptable range is \(c\in(\frac{2817}{2425})\) (given by Rivat and Sargos). ### Main Problems of the Paper This paper focuses on the application of the Piatetski - Shapiro sequence on dense sets. Specifically, the author studies the equation: \[ ab = [n^c] \] where \(a\in A\), \(b\in B\), \(n\in\mathbb{N}\), and \(c\in(1, 2)\), and \(A\) and \(B\) are dense sets. Rivat and Sárközy have proved that for dense sets \(A\) and \(B\) under certain conditions, when \(c\in(1, \frac{13}{11})\), the above equation can be solved. ### Main Contributions of the Paper In this paper, the author extends the effective range from \((1, \frac{13}{11})\) to \((1, \frac{6}{5})\) by using the method of three - dimensional exponential sums. Specifically, the author proves the following theorem: #### Theorem 1.1 There exist \(N_0 = N_0(c)>0\) and \(\delta=\delta(c)>0\) such that for \(N > N_0\) and any sets \(A, B\subseteq [N/2, N)\), if: \[ |A||B|\gg N^{2 - 2\delta} \] then the equation \(ab = [n^c]\) can be solved in the range of \(c\in(1, \frac{6}{5})\). In addition, the author also considers some special cases, such as when \(A = \mathbb{N}\) and \(B = Q_k\) (\(Q_k\) is the set of positive \(k\)-free integers), and obtains an asymptotic formula for the \(k\)-free divisor function. #### Theorem 1.2 Let \(\tau^{(k)}(n)\) be the \(k\)-free divisor function. Then for \(\gamma=\frac{1}{c}\) and \(c\in(1, \frac{6}{5})\), we have: \[ \sum_{1\leq n\leq x}\tau^{(k)}([n^c]) = cx\log x+\left(\frac{2\gamma_0 - 1}{\zeta(k)}-\frac{k\zeta'(k)}{\zeta^2(k)}+1 - c\right)x+O(x^{1 - \epsilon}) \] where \(\gamma_0\) is the Euler constant, \(\zeta(s)\) is the Riemann \(\zeta\) function, and \(\epsilon\) is an arbitrarily small positive number. ### Conclusion Through these results, the paper significantly expands the research scope of the Piatetski - Shapiro sequence and provides new tools and methods for subsequent research.