8
$\begingroup$

Let $F$ be a finite field of order $p$, where $p$ is prime. For any $n\times n$ matrix $A$ that is invertible over $F$, then there would appear to exist integers $k$ such that $A^{k} = I$. My question is simply what is the minimum value of $k$ (as a function of $n$ and $p$), which I will call $k_{n,p}$, such that this relation holds for all choices of $A$.

Proving that such an integer exists is relatively straightforward. By assumption, there exists a $B$ such that $AB = I$. Furthermore, there are precisely $\gamma_{n}(p) = (p^n - 1)(p^n-p) \ldots (p^n - p^{n-1})$ matrices satisfying the requirements for $A$. Since invertible matrices form a group, this implies that there exist distinct integers $a$ and $b$ between $1$ and $\gamma_{n}(p)+1$ such that $A^a = A^b$, and hence $B = A^{a-b-1} = A^{k-1}$. This implies that $k_{n,p} \leq \gamma_n(p)$. Furthermore, since the order of the group of invertible matrices over $F$ is $\gamma_n(p)$, and the powers of $A$ form a subgroup of this group, Lagrange's theorem implies that $k_{n,p}$ must divide $\gamma_n(p)$. So my question amounts to whether $k_{n,p} = \gamma_n(p)$, or whether you can do better.

This would essentially be an equivalent of Fermat's little theorem for matrices, but in searching such literature I have only been able to turn up trace relations, and so if this is well known, or if my reasoning is flawed, I would appreciate a pointer to the relevant literature (or to my error).

$\endgroup$
2
  • 2
    $\begingroup$ Do you want $A^k$ to be $A$ or $I$ mod p? The title says $A$ but the text of the question says $I$ which is more natural. (The questions are equivalent, with answers that differ by $1$.) $\endgroup$ Feb 23, 2019 at 15:49
  • $\begingroup$ I had intended to ask the former, but when writing the question realised that it was cleaner to ask the later as there would be less -1s floating about, and unfortunately forgot to update the title. $\endgroup$ Feb 24, 2019 at 2:40

1 Answer 1

9
$\begingroup$

The quantity that you are trying to compute is called the exponent of the group $\text{GL}_n(\mathbb F_p)$. As you note by Lagrange ,it always divides the order of the group. More generally, one can ask about the exponent of the group $\text{GL}_n(\mathbb F_q)$, where $q$ is a power of $p$. You'll find a detailed answer/discussion at https://math.stackexchange.com/questions/292620/exponent-of-gln-q

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.