Poissonian pair correlation for $αn^θ$

Maksym Radziwiłł,Andrei Shubin

2023-04-10

Abstract:We show that sequences of the form $\alpha n^{\theta} \pmod{1}$ with $\alpha > 0$ and $0 < \theta < \tfrac{43}{117} = \tfrac{1}{3} + 0.0341 \ldots$ have Poissonian pair correlation. This improves upon the previous result by Lutsko, Sourmelidis, and Technau, where this was established for $\alpha > 0$ and $0 < \theta < \tfrac{14}{41} = \tfrac{1}{3} + 0.0081 \ldots$. We reduce the problem of establishing Poissonian pair correlation to a counting problem using a form of amplification and the Bombieri-Iwaniec double large sieve. The counting problem is then resolved non-optimally by appealing to the bounds of Robert-Sargos and (Fouvry-Iwaniec-)Cao-Zhai. The exponent $\theta = \tfrac{2}{5}$ is the limit of our approach.

Number Theory

What problem does this paper attempt to address?

The problem that this paper attempts to solve is whether the sequence of the specific form \(\alpha n^\theta \mod 1\) has Poisson Pair Correlation (PPC). Specifically, the authors focus on whether these sequences exhibit Poisson Pair Correlation characteristics when \(\alpha > 0\) and \(0 < \theta < \frac{43}{117}\approx0.3675\).

### Background and Problem Statement

**Poisson Pair Correlation** means that for a given sequence \(\{x_n\}\), if the spacing distribution between its adjacent elements is similar to the spacing distribution of a random point set, then the sequence is said to have Poisson Pair Correlation. Specifically, for the \(\alpha n^\theta \mod 1\) sequence, the authors are concerned with:

\[ R_2([-\frac{s}{N}, \frac{s}{N}], \{\alpha n^\theta \mod 1\}, N) := \frac{1}{N} \# \left\{ 1 \leq n \neq m \leq N : N \| \alpha n^\theta - \alpha m^\theta \| \leq s \right\} \]

Whether the above expression approaches \(2s\) as \(N \to \infty\). Here \(\| x \| \) represents the distance from \(x\) to the nearest integer.

### Research Progress

Previous research has proven that under the conditions of \(\alpha > 0\) and \(0 < \theta < \frac{14}{41}\approx0.3415\), this sequence has Poisson Pair Correlation. The main contribution of this paper is to extend the range of \(\theta\) to \(0 < \theta < \frac{43}{117}\approx0.3675\).

### Methods and Techniques

To verify this extension, the authors adopted the following methods:
1. **Exponential sum estimation**: By introducing exponential sums and using the Poisson summation formula to transform them into a more tractable form.
2. **Amplification technique**: Use the amplification technique to simplify the problem.
3. **Double large sieve method**: Apply the Bombieri - Iwaniec double large sieve method to handle complex summation problems.
4. **Boundary estimation**: Utilize the results of Robert - Sargos and (Fouvry - Iwaniec -)Cao - Zhai to non - optimally solve the counting problem.

### Conclusion

Finally, the authors proved that for all \(\alpha > 0\) and \(0 < \theta < \frac{43}{117}\approx0.3675\), the sequence \(\alpha n^\theta \mod 1\) has Poisson Pair Correlation. This significantly improves the previous results and provides a new direction for further research.

---

In summary, this paper aims to verify and extend whether the \(\alpha n^\theta \mod 1\) sequence has Poisson Pair Correlation characteristics in a wider range of \(\theta\), thereby deepening our understanding of the pseudo - random properties of such sequences.