7
$\begingroup$

I am looking for a proof of the equality in the title (where $q\in\mathbb N$, ($q\ge2$)). Does anyone know such a proof?

Thanks in advance

$\endgroup$
1
  • 19
    $\begingroup$ Since $q^n-1=\prod_{d\mid n} \Phi_d(q)$, where $\Phi_d$ is the $d$th cyclotomic polynomial, we have $lcm(q-1,\dots,q^n-1) = lcm(\Phi_1(q),\dots,\Phi_n(q))$. But it's easy to see via the Euclidean algorithm that $\gcd(q^i-1,q^j-1)=q^{\gcd(i,j)}-1$, which implies that the $\Phi_k(q)$ are almost totally relatively prime with one another. So the quantity in question is almost exactly $\prod_{k\le n} \Phi_k(q) \approx \prod_{k\le n} q^{\phi(k)} = q^{\sum_{k\le n} \phi(k)}$. That's where the $3n^2/\pi^2$ comes from. $\endgroup$ Jun 1, 2014 at 7:57

1 Answer 1

9
$\begingroup$

Niven, Fermat's theorem for matrices, Duke Math. J. 15 (1948) 823–826, MR0026672 (10,183e), proved, for any positive integers $a$ and $r$, $${\rm lcm}[x-1,x^2-1,\dots,x^r-1]_{x=a}={\rm lcm}[a-1,a^2-1,\dots,a^r-1]$$ Now follow Greg Martin's comment.

$\endgroup$

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.