Abstract:Let us consider the following matrix $B_n$. The columns of $B_n$ are indexed with $[n]=\{1,2,\dots,n\}$ and the rows are indexed with $[n]^3$. The row corresponding to $(x_1,x_2,x_3)\in [n]^3$ is given by $\sum_{i=1}^3 e_{x_i}$, where $e_1,e_2,\dots,e_n$ is the standard basis of $\mathbb{R}^{[n]}$. Let $A_n$ be random $n\times n$ submatrix of $B_n$, where the probability that we choose a submatrix $C$ is proportional to $|\det(C)|^2$.
Let $p\ge 5$ be a prime. We prove that the asymptotic distribution of the $p$-Sylow subgroup of the cokernel of $A_n$ is given by the Cohen-Lenstra heuristics.
Our result is motivated by the conjecture that the first homology group of a random two dimensional hypertree is also Cohen-Lenstra distributed.
What problem does this paper attempt to address?
### What problem does this paper attempt to solve?
This paper mainly explores the Cohen - Lenstra distribution problem of sparse matrices under a specific probability measure. Specifically, the author studies a special class of random sparse matrices \(A_n\), where the rows and columns are selected based on determinantal biasing. Each row of such matrices contains at most 3 non - zero elements.
#### Main problem description:
1. **Matrix structure and probability measure**:
- Define a matrix \(B_n\), whose columns are indexed by the set \([n]=\{1, 2,\ldots,n\}\) and rows are indexed by the set \([n]^3\). The value of each row \((x_1, x_2, x_3)\in[n]^3\) is \(\sum_{i = 1}^3e_{x_i}\), where \(e_1, e_2,\ldots,e_n\) are the standard basis of \(\mathbb{R}^n\).
- Define a random sub - matrix \(A_n\), which is an \(n\times n\) sub - matrix randomly selected from \(B_n\), and the selection probability is proportional to the square of the determinant of the sub - matrix, that is \(P(X_n = K)=\frac{|det(B_n[K])|^2}{det(B_n^T B_n)}\).
2. **Cohen - Lenstra distribution of cokernel**:
- Study the asymptotic distribution of \(p\)-Sylow subgroups in the cokernel of \(A_n\) (i.e., the quotient group \(\mathbb{Z}^n/RowSpace(A_n)\)).
- Prove that when \(p\geq5\), the asymptotic distribution of \(p\)-Sylow subgroups in the cokernel of \(A_n\) conforms to the Cohen - Lenstra heuristic distribution.
3. **Motivation and application**:
- The motivation of this research comes from a conjecture: the first homology group of two - dimensional random hyper - trees should also conform to the Cohen - Lenstra distribution.
- This result shows that even sparse matrices can exhibit Cohen - Lenstra behavior in some cases, which provides theoretical support for the above conjecture.
#### Summary of mathematical formulas:
- **Matrix \(B_n\)**:
\[
B_n[(x_1, x_2, x_3)]=\sum_{i = 1}^3e_{x_i}
\]
- **Probability measure**:
\[
P(X_n = K)=\frac{|det(B_n[K])|^2}{det(B_n^T B_n)}
\]
- **Definition of cokernel**:
\[
cok(A_n)=\mathbb{Z}^n/RowSpace(A_n)
\]
- **Cohen - Lenstra distribution theorem**:
\[
\lim_{n\rightarrow\infty}P\left(\bigoplus_{i = 1}^s\Gamma_{n,i}\cong\bigoplus_{i = 1}^sG_i\right)=\prod_{i = 1}^s\left(\frac{1}{|Aut(G_i)|}\prod_{j = 1}^{\infty}(1 - p_i^{-j})\right)
\]
Through these studies, the author shows that the cokernel of sparse random matrices under a specific probability measure can follow the Cohen - Lenstra heuristic distribution, thus providing a new perspective for further understanding similar phenomena in random matrices and graph theory.