Abstract:The paper studies the probability for a Galois group of a random polynomial to be $A_n$. We focus on the so-called large box model, where we choose the coefficients of the polynomial independently and uniformly from $\{-L,\ldots, L\}$. The state-of-the-art upper bound is $O(L^{-1})$, due to Bhargava. We conjecture a much stronger upper bound $L^{-n/2 +\epsilon}$, and that this bound is essentially sharp. We prove strong lower bounds both on this probability and on the related probability of the discriminant being a square.
What problem does this paper attempt to address?
### What problem does this paper attempt to solve?
This paper mainly studies the probability problem that the Galois group of a random polynomial is the alternating group \(A_n\). Specifically, the authors focus on the so - called "large box model", in which the coefficients of the polynomial are independently and uniformly selected integers from the interval \([-L, L]\).
#### Research background
In probabilistic Galois theory, researchers usually consider the arithmetic properties of a random polynomial \(f\) with integer coefficients. One of the basic arithmetic properties is whether \(f\) is irreducible over the field of rational numbers \(\mathbb{Q}[X]\). More generally, each polynomial \(f\) has a Galois group \(G_f\), which acts naturally on the roots of \(f\), and thus can be regarded as a subgroup of the symmetric group \(S_{\deg f}\) (when \(f\) is a separable polynomial, this subgroup is well - defined up to conjugation).
#### Main problems
In this paper, the authors pay special attention to the following two aspects:
1. **Probability estimation**: For a given degree \(n\) and coefficient range \(L\), they hope to find the upper and lower bounds of the probability that the Galois group of the polynomial is the alternating group \(A_n\).
2. **Conjecture proposal**: The authors propose a stronger upper - bound conjecture \(L^{-n/2+\epsilon}\), and believe that this upper bound is essentially tight (sharp). They also prove strong lower bounds for this probability and the probability that the related discriminant is a square.
#### Specific goals
- **Improve existing results**: Currently, the best upper bound is \(O(L^{-1})\) given by Bhargava. The authors hope to obtain more accurate estimates through their research.
- **Verify the conjecture**: Provide new evidence to support their proposed conjecture about the asymptotic behavior of \(\text{Prob}(G_f = A_n)\).
#### Mathematical expressions
According to the content of the article, we can represent the key formulas in Markdown format:
- **Bhargava's result**:
\[
\text{Prob}(G_f\neq S_n)=\text{Prob}(G_f = S_{n - 1})+\text{Prob}(G_f = A_n)+O(L^{-2})
\]
\[
\text{Prob}(G_f = A_n)=O(L^{-1}),\quad\text{for }n\geq10
\]
- **Authors' conjecture**:
\[
L^{-n/2+\epsilon}\gg\text{Prob}(G_f = A_n)\gg L^{-n/2},\quad\text{as }L\to\infty
\]
In summary, this paper aims to deeply explore the probability problem that the Galois group of a random polynomial is the alternating group, and propose a stronger asymptotic estimation conjecture, thereby promoting the development of probabilistic Galois theory.