Verify this result when ## n=10 ##

  • Thread starter Thread starter Math100
  • Start date Start date
AI Thread Summary
The discussion focuses on verifying the equation ## \sum_{d\mid n}\sigma_{2}(d)\sigma_{2}(\frac{n}{d})=\sum_{d\mid n}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{n}{d}) ## for the case when ## n=10 ##. Participants calculate the necessary sigma function values, noting that ## \sigma_{0}(10) = 4 ## and ## \sigma_{2}(10) = 130 ##. They express the need to understand the convolution-like nature of the formulas and mention the Möbius function as a key concept in number theory. Ultimately, the calculations confirm that both sides of the equation equal 520 when evaluated for the divisors of 10. The conversation highlights the importance of understanding divisor functions in mathematical proofs.
Math100
Messages
813
Reaction score
229
Homework Statement
Prove that ## \sum_{d\mid n}\sigma_{2}(d)\sigma_{2}(\frac{n}{d})=\sum_{d\mid n}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{n}{d}) ##, and verify this result when ## n=10 ##.
Relevant Equations
None.
Proof:

Observe that
\begin{align*}
\sum_{d\mid n}\sigma_{2}(d)\sigma_{2}(\frac{n}{d})&=(N^{2}\ast u)\ast (N^{2}\ast u) \text{ since } \sigma_{2}(n)=\sum_{d\mid n}d^{2}\cdot 1=N^{2}\ast u)\\
&=(u\ast N^{2})\ast (N^{2}\ast u)\\
&=u\ast (N^{2}\ast N^{2})\ast u\\
&=u\ast (N^{2}\sigma_{0})\ast u \text{ since } N^{2}\ast N^{2}=\sum_{d\mid n}d^{2}(\frac{n^{2}}{d^{2}})=n^{2}\sum_{d\mid n}1=N^{2}\sigma_{0} \\
&=N^{2}\sigma_{0}\ast (u\ast u) \text{ since } u\ast u=\sigma_{0} \\
&=N^{2}\sigma_{0}\ast \sigma_{0}\\
&=\sum_{d\mid n}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{n}{d}).\\
\end{align*}
Therefore, ## \sum_{d\mid n}\sigma_{2}(d)\sigma_{2}(\frac{n}{d})=\sum_{d\mid n}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{n}{d}) ##.
Now we will verify this result when ## n=10 ##.
Consider the divisors ## d\in\left \{ 1, 2, 5, 10 \right \} ##.
Observe that
\begin{align*}
\sum_{d\mid 10}\sigma_{2}(d)\sigma_{2}(\frac{10}{d})&\stackrel{?}{=}\sum_{d\mid 10}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{10}{d})
\end{align*}
\begin{align*}
&\sigma_{2}(1)\sigma_{2}(\frac{10}{1})+\sigma_{2}(2)\sigma_{2}(\frac{10}{2})+\sigma_{2}(5)\sigma_{2}(\frac{10}{5})+\sigma_{2}(10)\sigma_{2}(\frac{10}{10})\\
&\stackrel{?}{=}1^{2}\sigma_{0}(1)\sigma_{0}(\frac{10}{1})+2^{2}\sigma_{0}(2)\sigma_{0}(\frac{10}{2})+5^{2}\sigma_{0}(5)\sigma_{0}(\frac{10}{5})+10^{2}\sigma_{0}(10)\sigma_{0}(\frac{10}{10}).
\end{align*}
 
Last edited by a moderator:
Physics news on Phys.org
I was thinking ## \sigma_{0}(1)=1, \sigma_{0}(2)=2, \sigma_{0}(5)=2, \sigma_{0}(10)=4 ##. But what should ## \sigma_{2}(1), \sigma_{2}(2), \sigma_{2}(5), \sigma_{2}(10) ## be?
 
Math100 said:
Homework Statement:: Prove that ## \sum_{d\mid n}\sigma_{2}(d)\sigma_{2}(\frac{n}{d})=\sum_{d\mid n}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{n}{d}) ##, and verify this result when ## n=10 ##.
Relevant Equations:: None.

Proof:

Observe that
\begin{align*}
\sum_{d\mid n}\sigma_{2}(d)\sigma_{2}(\frac{n}{d})&=(N^{2}\ast u)\ast (N^{2}\ast u) \text{ since } \sigma_{2}(n)=\sum_{d\mid n}d^{2}\cdot 1=N^{2}\ast u)\\
&=(u\ast N^{2})\ast (N^{2}\ast u)\\
&=u\ast (N^{2}\ast N^{2})\ast u\\
&=u\ast (N^{2}\sigma_{0})\ast u \text{ since } N^{2}\ast N^{2}=\sum_{d\mid n}d^{2}(\frac{n^{2}}{d^{2}})=n^{2}\sum_{d\mid n}1=N^{2}\sigma_{0} \\
&=N^{2}\sigma_{0}\ast (u\ast u) \text{ since } u\ast u=\sigma_{0} \\
&=N^{2}\sigma_{0}\ast \sigma_{0}\\
&=\sum_{d\mid n}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{n}{d}).\\
\end{align*}
Therefore, ## \sum_{d\mid n}\sigma_{2}(d)\sigma_{2}(\frac{n}{d})=\sum_{d\mid n}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{n}{d}) ##.
Now we will verify this result when ## n=10 ##.
Consider the divisors ## d\in\left \{ 1, 2, 5, 10 \right \} ##.
Observe that
\begin{align*}
\sum_{d\mid 10}\sigma_{2}(d)\sigma_{2}(\frac{10}{d})&\stackrel{?}{=}\sum_{d\mid 10}d^{2}\sigma_{0}(d)\sigma_{0}(\frac{10}{d})
\end{align*}
\begin{align*}
&\sigma_{2}(1)\sigma_{2}(\frac{10}{1})+\sigma_{2}(2)\sigma_{2}(\frac{10}{2})+\sigma_{2}(5)\sigma_{2}(\frac{10}{5})+\sigma_{2}(10)\sigma_{2}(\frac{10}{10})\\
&\stackrel{?}{=}1^{2}\sigma_{0}(1)\sigma_{0}(\frac{10}{1})+2^{2}\sigma_{0}(2)\sigma_{0}(\frac{10}{2})+5^{2}\sigma_{0}(5)\sigma_{0}(\frac{10}{5})+10^{2}\sigma_{0}(10)\sigma_{0}(\frac{10}{10}).
\end{align*}
I corrected the LaTeX. I cannot check the proof since I still don't know what ##N## and ##u## are.

The equation where I added a question mark needs to be checked, i.e. you have to calculate the sigma function values and add them. We have
$$
\sigma_k(n) =\sum_{d|n}d^k
$$
This means that ##\sigma_0## is the number of divisors, e.g. ##\sigma_0(10)= |\{1,2,5,10\}|=4.## For ##\sigma_2## we have to calculate e.g.
$$
\sigma_2(10) =\sum_{d|10}d^2=\sum_{d\in \{1,2,5,10\}} d^2 =1^2+2^2+5^2+10^2=130
$$

Or you look up the values in the table here:
https://de.wikipedia.org/wiki/Teilerfunktion
 
  • Like
Likes Math100
fresh_42 said:
I corrected the LaTeX. I cannot check the proof since I still don't know what ##N## and ##u## are.

The equation where I added a question mark needs to be checked, i.e. you have to calculate the sigma function values and add them. We have
$$
\sigma_k(n) =\sum_{d|n}d^k
$$
This means that ##\sigma_0## is the number of divisors, e.g. ##\sigma_0(10)= |\{1,2,5,10\}|=4.## For ##\sigma_2## we have to calculate e.g.
$$
\sigma_2(10) =\sum_{d|10}d^2=\sum_{d\in \{1,2,5,10\}} d^2 =1^2+2^2+5^2+10^2=130
$$

Or you look up the values in the table here:
https://de.wikipedia.org/wiki/Teilerfunktion
It's okay, you don't necessarily need to check the proof. I just wanted to know how to compute those values at the end, the verification part.
 
Math100 said:
It's okay, you don't necessarily need to check the proof. I just wanted to know how to compute those values at the end, the verification part.
The formula looks like a convolution, a variation of
$$
(f\ast g)(n)=\sum_{k\in D} f(k) g(n/k).
$$
Would be nice to understand the details, i.e. to see the proof why the divisor formulas can be written that way.
 
  • Like
Likes Math100
I think I have found it in a book. Möbius function seems to be the key word.
 
  • Informative
Likes Math100
fresh_42 said:
I think I have found it in a book. Möbius function seems to be the key word.
What book?
 
Math100 said:
What book?
A book on number theory. It's been a long time since I last time had a look.
 
fresh_42 said:
A book on number theory. It's been a long time since I last time had a look.
I see. And the results when ## n=10 ## match each other in the equation. It's ## 130+130+130+130=4+16+100+400=520 ##.
 
  • Like
Likes fresh_42
Back
Top