Abstract:A Steinhaus random multiplicative function $f$ is a completely multiplicative function obtained by setting its values on primes $f(p)$ to be independent random variables distributed uniformly on the unit circle. Recent work of Harper shows that $\sum_{n\le N} f(n)$ exhibits ``more than square-root cancellation," and in particular $\frac 1{\sqrt{N}} \sum_{n\le N} f(n)$ does not have a (complex) Gaussian distribution. This paper studies $\sum_{n\in {\mathcal A}} f(n)$, where ${\mathcal A}$ is a subset of the integers in $[1,N]$, and produces several new examples of sets ${\mathcal A}$ where a central limit theorem can be established. We also consider more general sums such as $\sum_{n\le N} f(n) e^{2\pi i n\theta}$, where we show that a central limit theorem holds for any irrational $\theta$ that does not have extremely good Diophantine approximations.
What problem does this paper attempt to address?
### The problems the paper attempts to solve
This paper aims to explore the distribution problem of the partial sums of random multiplicative functions on specific subsets, especially whether these partial sums satisfy the central limit theorem. Specifically:
1. **Research background**:
- Random multiplicative functions are an important model in number theory. By studying the behavior of these functions, new perspectives can be provided for understanding other important number - theoretic functions (such as Dirichlet characters, Liouville functions, and Möbius functions).
- Two common models of random multiplicative functions include: the Steinhaus model (a completely multiplicative function whose values on prime numbers are independent random variables uniformly distributed on the unit circle) and the Rademacher model (a multiplicative function on square - free integers taking values of ±1).
2. **Existing achievements**:
- Harper's breakthrough result shows that for the Steinhaus and Rademacher models, the size of the partial sum \(\sum_{n\leq N}f(n)\) is usually \(o(\sqrt{N})\), that is, "beyond - square - root cancellation".
3. **Objectives of this paper**:
- The objective of this paper is to study whether the partial sum \(\sum_{n\in A}f(n)\) satisfies the central limit theorem when the random multiplicative function is restricted to certain subsets \(A\) of [1, N].
- The author presents several new examples, proving that on certain subsets \(A\), the partial sum does satisfy the central limit theorem and improving some existing results.
4. **Main contributions**:
- A simplified condition is proposed such that when the subset \(A\) satisfies certain conditions, the distribution of the partial sum \(\sum_{n\in A}f(n)\) is close to the standard complex Gaussian distribution.
- A more general weighted sum \(\sum_{n\leq N}a_{n}f(n)\) is studied, and it is proved that in some cases, the weighted sum also satisfies the central limit theorem.
- For sums of the form \(\sum_{n\leq N}f(n)e^{2\pi in\theta}\), it is proved that when \(\theta\) is some specific type of irrational number, this sum also satisfies the central limit theorem.
In summary, the problem that this paper attempts to solve is: on which subsets \(A\) does the partial sum \(\sum_{n\in A}f(n)\) of the random multiplicative function satisfy the central limit theorem? And by proposing new conditions and examples, the existing achievements in this field are extended.