2
$\begingroup$

That is to say $\forall k \in \mathbb{N}$, there exist infinitely many $n\in \mathbb{N}$ s.t. $\mu \left ( n+1 \right )=\mu \left ( n+2 \right )=\cdots=\mu \left ( n+k \right )$, where $\mu$ is the Möbius function.

I saw this problem in a number theory textbook and it appears in the exercises for the section talking about the Möbius inversion formula. I cannot come up with an effective idea to tackle it.

This is not a homework problem.

Many thanks for your help!

$\endgroup$
  • $\begingroup$ Try making all $k$ of them equal to $0$ by forcing each number to have a repeated prime factor. Use the Chinese Remainder Theorem. $\endgroup$ – Dylan Oct 31 '15 at 10:17
5
$\begingroup$

Let $p_1,p_2,\cdots,p_k$ be any distinct prime numbers.

Then by the Chinese Remainder Theorem, the system of congruences $$\begin{align} n + 1 & \equiv 0 \mod p_1^2 \\ n + 2 & \equiv 0 \mod p_2^2 \\ n + 3 & \equiv 0 \mod p_3^2 \\ \vdots \\ n + k & \equiv 0 \mod p_k^2 \end{align}$$

has infinitely many solutions $n$. For any such solution $n$, we have that $n + m$ is divisible by $p_m^2$ for $m=1,2,3,\cdots,k$, and so we have that $\mu(n+m)=0$ for $m=1,2,3\cdots,k$

We see that there are infinitely many natural numbers $n$ such that $$ \mu(n+1)=\mu(n+2)=\mu(n+3)=\cdots=\mu(n+k)=0$$


edit

In fact, if $k \geq 4$ and for some $n$ we have that $$ \mu(n+1)=\mu(n+2)=\mu(n+3)=\cdots=\mu(n+k)$$ then it must in fact be the case that $$ \mu(n+1)=\mu(n+2)=\mu(n+3)=\cdots=\mu(n+k)=0$$

This is because amongst the set of numbers $$\lbrace n+1, n+2, n+3, n+4\rbrace$$ there will be a multiple of $4$. The Möbius function evaluated at this multiple of $4$ will be $0$.

$\endgroup$
  • $\begingroup$ I see. Great idea to apply CRT here. Thanks! $\endgroup$ – S. D. ZHU Oct 31 '15 at 12:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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