Prove Summation of Mobius Inversion w/ Sigma Function

  • Context: Graduate 
  • Thread starter Thread starter math_grl
  • Start date Start date
  • Tags Tags
    Inversion
Click For Summary

Discussion Overview

The discussion revolves around proving the summation involving the Möbius function and the divisor function, specifically \(\sum_{d \mid n} \mu(d) \sigma_0(d) = (-1)^{\omega(n)}\). Participants explore the definitions of the functions involved and the implications of their properties in the context of number theory.

Discussion Character

  • Technical explanation
  • Mathematical reasoning
  • Debate/contested

Main Points Raised

  • One participant seeks to prove the summation involving the Möbius function and the divisor function, questioning how \(F(n/d) = \sigma_0(d)\) fits into the proof.
  • Another participant requests clarification on the definitions of \(\omega\) and \(\sigma_0\), suggesting that \(\sigma_0(d) = F(n/d)\) might be a foundational definition.
  • A participant defines \(\sigma_0(n) = \sum_{d \mid n} 1\) and \(\omega(n) = \sum_{p \mid n} 1\), where \(p\) represents prime factors of \(n\).
  • One participant proposes a proof sketch, indicating that the sum's value for a composite \(n\) can be reduced to that of a squarefree \(n\), due to the properties of the Möbius function.
  • Another participant expresses intrigue at the proposed proof method, noting that they expected a more direct connection to Möbius inversion.
  • A later reply mentions a related summation \(\sum_{d \mid n} \mu(\frac{n}{d}) \sigma_0(d) = 1\), which may be relevant to the discussion.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the proof or the relationships between the functions. Multiple interpretations and approaches are presented, indicating ongoing debate and exploration of the topic.

Contextual Notes

Some definitions and assumptions regarding the functions \(\sigma_0\) and \(\omega\) are discussed, but their interrelations and implications remain partially unresolved. The discussion also highlights the complexity of proving the initial statement.

math_grl
Messages
46
Reaction score
0
Would like to show \sum_{d \mid n} \mu (d) \sigma_0 (d) = (-1)^{\omega (d)}.

This proof is just left out of text I'm looking at and I can't seem to piece how F(n/d) = \sigma_0 (d), where F(x) = \sum_{s \mid x} f(x).
 
Physics news on Phys.org
Hi,
can you make the problem a little clear? Who are omega and sigma_0, and how do they relate to little f or to big F. Maybe sigma_0(d) = F(n/d) is how sigma_0 is defined to begin with. Or not.
 
\sigma _0 (n) = \sum_{d \mid n} 1 and \omega (n) = \sum_{p \mid n} 1 where p is a prime and d is a divisor.

I thought the notation was standard for arithmetic functions.
 
Ok. I knew your sigma_0 as "tau", but never heard of omega.

I have a proof of the statement (I believe), but it is somewhat wordy. Here is a sketch:

First, realize that the value of that sum for n=(p1^whatever)(p2^whatever)(p3^whatever)... is the same as the value of the sum for another n, constructed as n=p1.p2.p3... (which is squarefree). This occurs because the mu() function will set to 0 the terms of the sum where the divisor d is not squarefree. So actually you only need to prove the statement for a squarefree n.

That said, the divisors of a squarefree n are just combinations of distinct primes; their corresponding mu values are either 1 or -1. And there will be two cases: one where all the mu(d) are equal to mu(n/d), and other when all the mu(d) are equal to minus mu(n/d). And these two cases occur precisely when omega(n) is even or odd, respectively.
 
Last edited:
Now you've given me something to think about...

very interesting way of proving it as I thought being in the same section as that of Mobius inversion, was supposed to follow direct from that or something...
 
It sort of follows... it is based on
<br /> \sum_{d \mid n} \mu (\frac n d) \sigma_0 (d) = 1<br />
which I forgot to mention in my previous post.
 

Similar threads

  • · Replies 8 ·
Replies
8
Views
2K
  • · Replies 1 ·
Replies
1
Views
3K
Replies
1
Views
4K
  • · Replies 52 ·
2
Replies
52
Views
4K
  • · Replies 13 ·
Replies
13
Views
2K
  • · Replies 7 ·
Replies
7
Views
2K
  • · Replies 31 ·
2
Replies
31
Views
3K
  • · Replies 25 ·
Replies
25
Views
4K
  • · Replies 15 ·
Replies
15
Views
5K
  • · Replies 5 ·
Replies
5
Views
2K