Abstract:We show that $0,1$-polynomials of high degree and few terms are irreducible with high probability. Formally, let $k\in\mathbb{N}$ and $F(x)=1+\sum_{i=1}^kx^{n_i}$, where $ 0<n_1<\cdots<n_k\leq N. $ Then we show that $\lim_{k\rightarrow\infty}\limsup_{N\rightarrow\infty}\mathbb{P}(\text{$F(x)$ is reducible})=0.$ The probability in this context is derived from the uniform count of polynomials $F(x)$ of the above form.
What problem does this paper attempt to address?
### What problem does this paper attempt to solve?
The main goal of this paper is to prove that sparse polynomials (lacunary polynomials) with 0,1 coefficients are almost surely irreducible when they are of high degree and have a small number of terms. Specifically, the author studied polynomials of the form \(F(x) = 1+\sum_{i = 1}^{k}x^{n_i}\), where \(0 < n_1<\cdots < n_k\leq N\), and showed the following conclusion:
\[
\lim_{k\rightarrow\infty}\limsup_{N\rightarrow\infty}P(F(x)\text{ is reducible}) = 0
\]
Here, the probability is uniformly selected from the polynomials \(F(x)\) of the above form.
#### Background and motivation of the main problem
1. **Abundance of irreducible polynomials**:
- In 1936, van der Waerden proved that for polynomials \(f(x)=x^k+\sum_{i = 0}^{k - 1}a_ix^i\) with a fixed degree \(k\) and randomly selected coefficients \(a_i\), when the range of coefficients increases, these polynomials are almost surely irreducible.
- Subsequent research focused on polynomial families with a fixed height and let the degree grow. For example, in 2020, Bary - Soroker and Kozma proved that if all coefficients \(a_i\in[1,210]\), then as the degree \(k\) approaches infinity, these polynomials are almost surely irreducible.
2. **Irreducibility of polynomials with 0,1 coefficients**:
- Odlyzko and Poonen proposed a conjecture in 1993: Almost all polynomials with 0,1 coefficients are irreducible. Although previous work has proven that some polynomials with 0,1 coefficients are irreducible, completely solving this conjecture still poses challenges.
3. **Irreducibility of sparse polynomials**:
- This paper specifically focuses on sparse polynomials with 0,1 coefficients (i.e., polynomials with a small number of terms). By introducing new methods, the author proved that such polynomials are almost surely irreducible when they are of high degree and have a small number of terms.
#### Main results
The core result of the article is to prove the following theorem:
\[
\text{Theorem 1:}\quad\lim_{k\rightarrow\infty}\limsup_{N\rightarrow\infty}P(F(x)\text{ is reducible}) = 0
\]
To prove this theorem, the author used two key facts:
1. **Existence of non - trivial non - cyclotomic factors**:
- Using existing results, it was proven that the number of polynomials with non - trivial non - cyclotomic factors is very small.
2. **Existence of cyclotomic factors**:
- It was proven that the number of polynomials with cyclotomic factors also tends to zero as \(k\) and \(N\) increase.
Through these two results, the author finally concluded that sparse 0,1 - coefficient polynomials are almost always irreducible.
### Summary
This paper solves the problem that sparse 0,1 - coefficient polynomials are almost always irreducible when they are of high degree and have a small number of terms by introducing new mathematical tools and techniques, providing further support for Odlyzko and Poonen's conjecture.