0

I'm working on a problem that states: Let $a, b, m, n$ be natural numbers, $a > 1$, and suppose that $a$ and $b$ have no common factors (GCD). We are asked to prove that if $a^m + b^m$ is divisible by $a^n + b^n$, then $m$ must be divisible by $n$.

I've tried using long division and modular arithmetic, but I'm stuck on how to proceed with the proof. The textbook solution simply suggests "carrying out long division," which I did without success.

After seeing the comments, one user pointed out that this problem $a,b,m,n$ be positive integers such that g.c.d.$(a,b)=1$ and $a^m+b^m\mid a^n+b^n$ , then $m\mid n$? offers the same solution.

And that is indeed similar. I have seen the answers, but I found them unhelpful. It assumed some prerequisite knowledge that I don't have, as of now, or perhaps, it skipped some steps or didn't give as elaborate of an explanation as my brain needed.

Okay, after seeing some comments, and answers, I have advanced my understanding of the problem.

So, I have assumed that $n <= m$, and I have used the division algorithm to express the following statement $m = nq + r$ for $0<=r<n$.

$a^n + b^n | a^m + b^m$. From this statement, I have thought of trying to do $a^m + b^m = (a^n + b^n)x$, and further try to solve the problem.

Can someone help me get started on the proof? Understanding the logic and intuition behind each step would be very helpful.

  • Welcome to MSE. Can you show us your work using the long division? We would then understand your difficulty. – eti902 Jul 20 '24 at 15:06
  • I tried to long divide a^m + b^m by a^n + b^n, as that is the first thing I saw. I didn't get any meaningful conclusions. – AshishMath Jul 20 '24 at 15:14
  • Try doing it with some numerical examples first (for example, fix $n=2$ and try $m=4,5,6$), and you should be able to find a pattern for the general case. – eti902 Jul 20 '24 at 15:22
  • if you use long division you get a succession of $a^{m-n}b^n, a^{n-2n}b^{2n},...a^{m-kn}$ and never $a^0 $ if n does not divide m. – trula Jul 20 '24 at 15:23
  • Also: https://math.stackexchange.com/q/3530503/42969 . – Martin R Jul 20 '24 at 15:39
  • I added a skipped step in the linked dupe. If it still is unclear then please say precisely what is not clear so we can elaborate further. $\ \ $ – Bill Dubuque Jul 20 '24 at 18:49

1 Answers1

0

Here is a sketch: First show that $n\leq m$. Then note that $a^m+b^m=a^{m-n}(a^n+b^n)-a^{m-n}b^n+b^{m-n}(a^n+b^n)-b^{m-n}a^n$ $=(a^{m-n}+b^{m-n})(a^n+b^n)-a^{i}b^{i}(a^{j}+b^{j})$, where $i=\min(n,m-n)$ and $j$ accordingly either $m-2n$ or $2n-m$. The assumptions on $a,b$ imply that $a^n+b^n$ divides $a^{j}+b^{j}$. If $j=m-n$, then $j< m$ and if $j=2n-m$, then since $n\leq m$ we have $m\geq 2n-m$. In both cases $j\leq m$. In case of equality, $n$ divides $m$. If $j<m$ use induction, by which $n$ divides $j$ and thus $m$ as well.

dialegou
  • 872
  • How is $a^{m-n}(a^n + b^n) - a^{m-n}b^n + b^{m-n}(a^n+b^n) - b^{m-n}a^n = (a^{m-n} + b^{m-n})(a^n + b^n) - a^{i}b^{i}(a^j + b^j)$ where $i = min(n, m-n)$ and j accordingly either $m-2n$ or $2n - m$. I am a beginner myself, in this field, and I don't understand this manipulation. – AshishMath Jul 20 '24 at 15:47
  • Can you be more specific on what you understand and what you don't understand about the manipulation? This way I don't know how to respond. – dialegou Jul 20 '24 at 17:12