Adam J. Harper
Abstract:We give a short review of recent progress on determining the order of magnitude of moments $\mathbb{E}|\sum_{n \leq x} f(n)|^{2q}$ of random multiplicative functions, and of closely related issues.
We hope this can serve as a concise introduction to some of the ideas involved, for those who may not have too much background in the area.
What problem does this paper attempt to address?
### Problems the paper attempts to solve
This paper aims to explore the magnitude of the moments of random multiplicative functions. Specifically, the author reviews the recent research progress on determining the absolute power moments \(E|\sum_{n \leq x} f(n)|^{2q}\) of the random multiplicative function \(\sum_{n \leq x} f(n)\), where \(f(n)\) is a Steinhaus random multiplicative function, that is, its values are uniformly distributed on the complex unit circle.
#### Main problems
1. **Estimation of low - order moments**: For \(0 \leq q < 1\), the author studies the asymptotic behavior of these moments and proposes accurate upper and lower bound estimates.
2. **Estimation of high - order moments**: For \(q \geq 1\), the author explores the growth rate of the moments, especially the behavior when \(q\) is large.
#### Key challenges
- **Influence of dependence**: There are dependencies among the values of random multiplicative functions (for example, \(f(6)=f(2)f(3)\)), which makes their behavior different from the classical probability setting of independent random variable sequences.
- **Complexity of moments**: Except for the simplest cases of \(q = 0\) and \(q = 1\), moment estimation in other cases is very complex and involves sophisticated probability and number - theoretic analysis.
#### Main results
- **Results for low - order moments**: For \(0 \leq q \leq 1\), the author proves that:
\[
E\left|\sum_{n \leq x} f(n)\right|^{2q} \asymp \left(x^{1+(1 - q)\sqrt{\log\log x}}\right)^q.
\]
- **Results for high - order moments**: For \(1 \leq q \leq c\frac{\log x}{\log\log x}\), the author gives more complex upper and lower bound estimates:
\[
e^{-q^{2}\log q - q^{2}\log\log(2q)-Cq^{2}}\leq\frac{E\left|\sum_{n \leq x} f(n)\right|^{2q}}{x^{q}(\log(q - 1)^{2}x)}\leq e^{-q^{2}\log q - q^{2}\log\log(2q)+Cq^{2}}.
\]
Through these results, the author not only solves the Helson conjecture (that is, the first absolute moment should be \(o(\sqrt{x})\)), but also reveals the subtle behavior of the moments of random multiplicative functions at different \(q\) values, especially the subdiffusivity which is better than the square root in low - order moments.