MHB Find all functions satisfying f(mn)=f(m)f(n), and m+n|f(m)+f(n)

  • Thread starter Thread starter lfdahl
  • Start date Start date
  • Tags Tags
    Functions
Click For Summary
The discussion focuses on finding functions f: ℕ → ℕ that satisfy the equations f(mn) = f(m)f(n) and m+n | f(m) + f(n) for all natural numbers m and n. Participants note that odd powers of n exhibit this property, suggesting a relationship between the equations and polynomial forms. There is uncertainty about whether the identified solutions are exhaustive. A typo in the suggested solution is also mentioned, indicating a need for clarification. The conversation emphasizes the exploration of functional equations within number theory.
lfdahl
Gold Member
MHB
Messages
747
Reaction score
0
Find all functions, $f: \mathbb{N}\rightarrow \mathbb{N}$, satisfying

\[f(mn)=f(m)f(n),\: \: \: and \: \: \: m+n \: \: |\: \: f(m)+f(n)\]for all $m,n \in \mathbb{N}$.
 
Mathematics news on Phys.org
lfdahl said:
Find all functions, $f: \mathbb{N}\rightarrow \mathbb{N}$, satisfying

\[f(mn)=f(m)f(n),\: \: \: and \: \: \: m+n \: \: |\: \: f(m)+f(n)\]for all $m,n \in \mathbb{N}$.

by putting m = 1 we get $f(n) = f(1) f(n)$ so $f(1) = 1$

now 1 + n is a factor of 1 + f(n) so by inspection f(n) = n. I do not have a proof of the same
 
Last edited:
kaliprasad said:
by putting m = 1 we get $f(n) = f(1) f(n)$ so $f(1) = 1$

now 1 + n is a factor of 1 + f(n) so by inspection f(n) = n. I do not have a proof of the same
[sp]Odd powers of $n$ also have this property: $(mn)^k = m^kn^k$, and $(m+n)^k = (m+n)(m^{k-1} - m^{k-2}n + \ldots + n^k)$ if $k$ is an odd integer.

I have not thought about whether these are the only solutions.
[/sp]
 
Opalg said:
[sp]Odd powers of $n$ also have this property: $(mn)^k = m^kn^k$, and $(m+n)^k = (m+n)(m^{k-1} - m^{k-2}n + \ldots + n^k)$ if $k$ is an odd integer.

I have not thought about whether these are the only solutions.
[/sp]

Right
 
Opalg said:
[sp]Odd powers of $n$ also have this property: $(mn)^k = m^kn^k$, and $(m+n)^k = (m+n)(m^{k-1} - m^{k-2}n + \ldots + n^k)$ if $k$ is an odd integer.

I have not thought about whether these are the only solutions.
[/sp]

there is a typo in above

$(m+n)^k = (m+n)(m^{k-1} - m^{k-2}n + \ldots + n^k)$

should be

$m^k+n^k = (m+n)(m^{k-1} - m^{k-2}n + \ldots + n^{k-1})$
 
Suggested solution:
As kaliprasad noted: $f(1) = 1$. So, $2n+1 \;\; | \;\; f(2n) + f(1) = f(2) \cdot f(n) + 1 \Rightarrow gcd(2n+1,f(2)) = 1$ for all $n$. This means, that $f(2)$ has no odd prime divisor, so $f(2) = 2^k$. Furthermore, $3 \;\; | 1 + f(2) = 1 + 2^k \Rightarrow k$ is odd (as Opalg noted). Also note, $f(2^m) = f(2) \cdot f(2) \cdot … \cdot f(2) = 2^{km}$ for all $m$. Now, for all $m$ and $n$, $2^m+n \;\; | \;\; f(2^{m}) + f(n) = 2^{km} + f(n)$. But $k$ is odd, so $2^m+n \;\; | \;\; 2^{km} + n^k$. Thus $2^m+n \;\; | \;\; f(n)-n^k$. But this is valid for all $m$, so $f(n) – n^k = 0 \Rightarrow f(n) = n^k.$
We have now proven, that $f(n) = n^k$ for all $n$ for some fixed odd number $k$. On the other hand, it is easily seen that functions of this form satisfy the given conditions.
 
Insights auto threads is broken atm, so I'm manually creating these for new Insight articles. In Dirac’s Principles of Quantum Mechanics published in 1930 he introduced a “convenient notation” he referred to as a “delta function” which he treated as a continuum analog to the discrete Kronecker delta. The Kronecker delta is simply the indexed components of the identity operator in matrix algebra Source: https://www.physicsforums.com/insights/what-exactly-is-diracs-delta-function/ by...

Similar threads

  • · Replies 4 ·
Replies
4
Views
1K
  • · Replies 1 ·
Replies
1
Views
1K
Replies
3
Views
1K
  • · Replies 1 ·
Replies
1
Views
2K
  • · Replies 6 ·
Replies
6
Views
1K
  • · Replies 3 ·
Replies
3
Views
2K
  • · Replies 9 ·
Replies
9
Views
2K
  • · Replies 2 ·
Replies
2
Views
1K
  • · Replies 4 ·
Replies
4
Views
2K
Replies
10
Views
2K