1

Lately I have been studying the divisibility of some cyclic sums, and I was wondering about the following

Conjecture

Let it be a set of distinct positive integers $S=\{x_1,x_2,...,x_n\}$ such that $2\leq{x_1}<x_2<...<x_n$. Then,

$$\prod_{k=1}^{n}x_{k}\nmid\left(\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^nx_k\right)$$

I would appreciate any help regarding this conjecture proof or refutation.

I already noted that there exist sets of distinct integers $S=\{x_1,x_2,...,x_n\}$ such that $$\prod_{k=1}^{n}x_{k}\mid\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)$$

For example, for $S=\{2,3,6\}$, $$2*3*6\mid (2*3)+(3*6)+(6*2)$$

Thanks in advance!

EDIT

I found that the conjecture is true if $\min\{x_1,x_2,...,x_n\}>n$, as it follows from the following

Proof

Assuming $x_{1}=x_{2}=...=x_{n}=n+1$, we get that

$$\prod_{k=1}^{n}x_{k}=\left(n+1\right)^{n}$$

$$\left(\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^{n}x_{k}\right)=n\left(\left(n+1\right)^{n-1}+\left(n+1\right)^{n-2}+...+\left(n+1\right)\right)$$

$$n\left(\left(n+1\right)^{n-1}+\left(n+1\right)^{n-2}+...+\left(n+1\right)\right)=\left(\left(n+1\right)-1\right)\left(\left(n+1\right)^{n-1}+\left(n+1\right)^{n-2}+...+\left(n+1\right)\right)$$

$$\left(\left(n+1\right)-1\right)\left(\left(n+1\right)^{n-1}+\left(n+1\right)^{n-2}+...+\left(n+1\right)\right)=\left(n+1\right)^{n}-\left(n+1\right)$$

Subsequently,

$$\left(\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^{n}x_{k}\right)=\left(n+1\right)^{n}-\left(n+1\right)$$

Thus,

$$\prod_{k=1}^{n}x_{k}>\left(\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^{n}x_{k}\right)$$

And therefore,

$$\prod_{k=1}^{n}x_{k}\nmid\left(\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^{n}x_{k}\right)$$

Subsequently, the conjecture is true if $\min\{x_1,x_2,...,x_n\}>n$.

Juan Moreno
  • 1,110

1 Answers1

1

The conjecture is true for $n=2$ since $$\small x_1+x_2=kx_1x_2\implies (kx_1-1)(kx_2-1)=1\implies kx_1-1=kx_2-1\implies x_1=x_2$$


The conjecture is false for $n=3$ since for $(x_1,x_2,x_3)=(2,4,14)$, we have$$\frac{x_1x_2+x_2x_3+x_3x_1+x_1+x_2+x_3}{x_1x_2x_3}=1$$


The conjecture is false for every $n\ge 3$.

Proof :

Let $$S(x_1,x_2,\cdots,x_n):=\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^nx_k$$

We have $$S(x_1,x_2,\cdots,x_n)=S(x_1,x_2,\cdots,x_{n-1})+x_nS(x_1,x_2,\cdots,x_{n-1})+x_n+\prod^{n-1}_{k=1}x_k$$

So, taking $$x_n=S(x_1,x_2,\cdots,x_{n-1})+\prod^{n-1}_{k=1}x_k$$ we get $$\frac{S(x_1,x_2,\cdots,x_n)}{\displaystyle\prod_{k=1}^{n}x_k}=\frac{2+S(x_1,x_2,\cdots,x_{n-1})}{\displaystyle\prod_{k=1}^{n-1}x_k}=\frac{2+S(x_1,x_2,\cdots,x_{n-2})+x_{n-1}S(x_1,x_2,\cdots,x_{n-2})+x_{n-1}+\displaystyle\prod^{n-2}_{k=1}x_k}{\displaystyle\prod_{k=1}^{n-1}x_k}$$

Now, taking $$x_{n-1}=2+S(x_1,x_2,\cdots,x_{n-2})+\prod^{n-2}_{k=1}x_k$$ we get $$\frac{S(x_1,x_2,\cdots,x_n)}{\displaystyle\prod_{k=1}^{n}x_k}=\frac{2+S(x_1,x_2,\cdots,x_{n-1})}{\displaystyle\prod_{k=1}^{n-1}x_k}=\frac{2+S(x_1,x_2,\cdots,x_{n-2})}{\displaystyle\prod_{k=1}^{n-2}x_k}$$

Similarly, taking $$x_{i}=2+S(x_1,x_2,\cdots,x_{i-1})+\prod^{i-1}_{k=1}x_k$$ for $i=3,4,\cdots,n-2$, we get $$\frac{S(x_1,x_2,\cdots,x_n)}{\displaystyle\prod_{k=1}^{n}x_k}=\frac{2+S(x_1,x_2)}{\displaystyle\prod_{k=1}^{2}x_k}=\frac{2+x_1+x_2}{x_1x_2}$$ which equals $1$ when $(x_1,x_2)=(2,4)$.

So, a counterexample is $$x_i=\begin{cases}S(x_1,x_2,\cdots,x_{n-1})+\displaystyle\prod^{n-1}_{k=1}x_k&\ \ \text{if $\ \ i=n$} \\\\2+S(x_1,x_2,\cdots,x_{i-1})+\displaystyle\prod^{i-1}_{k=1}x_k&\ \ \text{if $\ \ 3\le i\le n-1$} \\\\4&\ \ \text{if $\ \ i=2$}\\\\2&\ \ \text{if $\ \ i=1$}\end{cases}$$ where $$S(x_1,x_2,\cdots,x_n)=\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^nx_k$$

For example,

  • $(x_1,x_2,x_3,x_4)=(2,4,16,254)$ is a counterexample for $n=4$.

  • $(x_1,x_2,x_3,x_4,x_5)=(2,4,16,256,65534)$ is a counterexample for $n=5$.$\quad\blacksquare$


Added :

Let $$S(x_1,x_2,\cdots,x_n):=\sum_{cyc}\left(\prod_{k=1}^{n-1}x_{k}\right)+\sum_{cyc}\left(\prod_{k=1}^{n-2}x_{k}\right)+...+\sum_{k=1}^nx_k$$

Let us call your conjecture Conjecture 1.

Conjecture 1 : If $x_1,x_2,\cdots,x_n$ are positive integers satisfying $2\le x_1\lt x_2\lt\cdots\lt x_n$, then $\displaystyle\prod_{k=1}^{n}x_k\not\mid S(x_1,x_2,\cdots,x_n)$.

We've already seen that Conjecture 1 is true for $n=2$, and is false for $n\ge 3$.

Now, let us consider the following conjecture :

Conjecture 2 : If $x_1,x_2,\cdots,x_n$ are positive integers satisfying $2\le x_1\lt x_2\lt\cdots\lt x_n$ and $\color{red}{\gcd(x_1,x_2,\cdots,x_n)=1}$, then $\displaystyle\prod_{k=1}^{n}x_k\not\mid S(x_1,x_2,\cdots,x_n)$.

Conjecture 2 is true for $n=2$ since Conjecture 1 is true for $n=2$.

In the following, I'm going to prove the following claim :

Claim : Conjecture 2 is true for $n=3,4$.

Proof :

Let $A:=S(x_1,x_2,x_3)$.

Suppose that there are positive integers $x_1,x_2,x_3$ satisfying $2\le x_1\lt x_2\lt x_3, \gcd(x_1,x_2,x_3)=1$ and $x_1x_2x_3\mid A$.

If exactly one of $x_1,x_2,x_3$ is even, then $A$ is odd, so $x_1x_2x_3\not\mid A$.

If exactly one of $x_1,x_2,x_3$ is odd, then $A$ is odd, so $x_1x_2x_3\not\mid A$.

So, we see that $x_1,x_2,x_3$ have to be odd to have $$x_1x_2x_3\mid A\implies x_1x_2x_3\mid \frac A2\implies \frac A2\ge x_1x_2x_3\implies 4A\ge 8x_1x_2x_3$$

$$\implies X_1X_2X_3\le 3(X_1+X_2+X_3)+10$$ $$\implies X_1X_2\le 3\bigg(\frac{X_1}{X_3}+\frac{X_2}{X_3}+1\bigg)+\frac{10}{X_3}\lt 3\times 3+\frac{10}{13}\lt 10\tag1$$ where $X_i:=2x_i-1$ for $i=1,2,3$.

Also, since $x_1\ge 3$ and $x_2\ge 5$, we have $$X_1X_2=(2x_1-1)(2x_2-1)\ge 5\times 9=45\tag2$$

There is no $(x_1,x_2)$ satisfying both $(1)$ and $(2)$.

So, Conjecture 2 is true for $n=3$.

Next, let $B:=S(x_1,x_2,x_3,x_4)$.

Suppose that there are positive integers $x_1,x_2,x_3,x_4$ satisfying $2\le x_1\lt x_2\lt x_3\lt x_4, \gcd(x_1,x_2,x_3,x_4)=1$ and $x_1x_2x_3x_4\mid B$.

If exactly one of $x_1,x_2,x_3,x_4$ is even, then $B$ is odd, so $x_1x_2x_3x_4\not\mid B$.

If exactly two of $x_1,x_2,x_3,x_4$ are even, then $B$ is odd, so $x_1x_2x_3x_4\not\mid B$.

If exactly three of $x_1,x_2,x_3,x_4$ is odd, then $B$ is odd, so $x_1x_2x_3x_4\not\mid B$.

So, we see that $x_1,x_2,x_3,x_4$ have to be odd to have $$x_1x_2x_3x_4\mid B\implies x_1x_2x_3x_4\mid \frac B2\implies \frac B2\ge x_1x_2x_3x_4\implies 8B\ge 16x_1x_2x_3x_4$$

$$\implies X_1X_2X_3X_4\le 3(X_1X_2+X_1X_3+X_1X_4+X_2X_3+X_2X_4+X_3X_4)+12(X_1+X_2+X_3+X_4)+31$$ $$\implies X_1X_2\le 3\bigg(\frac{X_1X_2}{X_3X_4}+\frac{X_1}{X_4}+\frac{X_1}{X_3}+\frac{X_2}{X_4}+\frac{X_2}{X_3}+1\bigg)+12\bigg(\frac{X_1}{X_3X_4}+\frac{X_2}{X_3X_4}+\frac{1}{X_3}+\frac{1}{X_4}\bigg)+\frac{31}{X_3X_4}$$ $$\lt 3\times 6+12\bigg(2+\frac{1}{13}+\frac{1}{17}\bigg)+\frac{31}{13\times 17}=\frac{569}{13}\lt 44\tag3$$

where $X_i:=2x_i-1$ for $i=1,2,3,4$.

Also, since $x_1\ge 3$ and $x_2\ge 5$, we have $$X_1X_2=(2x_1-1)(2x_2-1)\ge 5\times 9=45\tag4$$

There is no $(x_1,x_2)$ satisfying both $(3)$ and $(4)$.

So, Conjecture 2 is true for $n=3,4$.$\quad\blacksquare$

mathlove
  • 139,939
  • Thanks @mathlove; unfortunately, when looking at your answer I realized that (i) it was intended to mean positive integers, and also (ii) $2\leq{x_1}<x_2<...<x_n$. Notwithstanding, thanks for your answer, which has highlighted this two errors in the question posing (already edited). If you could help with the rest of cases ($n>2$) it would be great! – Juan Moreno Nov 09 '20 at 12:13
  • 1
    @Juan Moreno : I've added a counterexample satisfying $2\le x_1\lt x_2\lt \cdots\lt x_n$ for $n=3$. – mathlove Nov 09 '20 at 12:44
  • Great @mathlove, thanks! For you the bounty. I see that the three numbers have some common factor greater than one...how would the conjecture be if we set the restriction of one of the positive integers of the set being pairwise coprime to the other ones? Do you have any clue? – Juan Moreno Nov 09 '20 at 15:49
  • 1
    @Juan Moreno : I've added a proof that the conjecture is false for every $n\ge 3$. I'll answer the question in your comment later if possible. – mathlove Nov 10 '20 at 07:35
  • 1
    superb!! I like a lot the proof because it is constructive. Many thanks!! – Juan Moreno Nov 10 '20 at 12:07
  • 1
    @Juan Moreno : I've added a proof that Conjecture 2 is true for $n=3,4$. I've spent a lot of time on the cases for $n\ge 5$, but with no success. So, I don't know if Conjecture 2 is true for $n\ge 5$. – mathlove Nov 14 '20 at 12:36
  • 1
    Really interesting! I will take a look and work a bit on it, although I doubt I will suceed on proving further cases. Many thanks @mathlove! – Juan Moreno Nov 14 '20 at 15:40