Is This Proof Valid? - $\sum_{k=1}^{n}\mu(k!)=1$ for $n \geq 3$

522 Views Asked by At

I decided to prove this by induction and I'm not sure if this is a valid proof, or maybe there is something that I'm missing. Anyway, I'd appreciate some feedback nonetheless.

Note: This proof assumes knowledge of the Möbius function.

Proof that $\sum_{k=1}^{n}\mu(k!)=1$ for $n \geq 3$:

Base Case

  • Let $n=3$ and test:

$\sum_{k=1}^{3}\mu(k!) = \mu(1) + \mu(2 \cdot 1) + \mu(3 \cdot 2 \cdot 1)$

$=(1)+(-1)+(1)=1$

  • True for base case ($n=3$).

Induction Hypothesis

  • Assume that it is true for $m$ for $m \geq 3$:

$\sum_{k=1}^{m}\mu(k!)=1$

Inductive Step:

  • Using the Inductive Hypothesis as a premise, we have:

$\sum_{k=1}^{m+1}\mu(k!)=(\sum_{k=1}^{m}\mu(k!)=1)+\mu((m+1)!)$

$=1+\mu((m+1)!)$

Since $m\geq 3$, we have $m+1\geq 4$. But then $(m+1)!=(m+1)(m)...(4)(3)(2)(1)$ Notice the factor of $(4)$ in the expanded sum.

Since $4=2^2$ is a square number, $(m+1)!$ has a squared factor, and according to the definition of the Möbius function, $\mu((m+1)!)=0$.

Finally, we have

$=1+\mu((m+1)!) = 1+0 = 1$

  • True for inductive step, therefore true for all cases.