9
$\begingroup$

Consider four positive numbers (not necessarily integers). The pairwise products are $2$, $3$, $4$, $5$, $6$, plus one more number.

What is the 6th product? What are the numbers?

I found this from Quora and I would be interested in a nice solution!

If we name the four numbers $x_1, x_2, x_3, x_4$ and the missing product $p_6$, then all of the possible products are:

$$x_1 x_2,\quad x_1 x_3,\quad x_1 x_4,\quad x_2 x_3,\quad x_2 x_4,\quad\text{and}\quad x_3 x_4$$

There are six equations and five unknowns, but I don't know how to assign the six different numbers to each of them.

I understand that the partial products which do not share a common factor (for example, $x_1 x_2$ and $x_3 x_4$) should not be assigned to numbers which do have a common factor, for example $2$ and $4$, or $2$ and $6$, or $3$ and $6$.

$\endgroup$
7
  • 2
    $\begingroup$ Yes, I've seen this myself - I don't think any of the replies there is correct! $\endgroup$ Apr 7 '20 at 13:01
  • 3
    $\begingroup$ Are these meant to be integers? But: The only way to get a prime $p$ as a product is $1\times p$. Thus, since $2,3,5$ are on the list of products, we need to have $1,2,3,5$ on the list of numbers, hence, that's the full list. Which does not work since $10,15$ are not among the products. $\endgroup$
    – lulu
    Apr 7 '20 at 13:08
  • 1
    $\begingroup$ What is the source of the problem? I expect some information has been omitted. If it's online, please provide a link. $\endgroup$
    – lulu
    Apr 7 '20 at 13:10
  • $\begingroup$ @lulu good point, not necessarily integers. I edited my question. $\endgroup$ Apr 7 '20 at 13:10
  • 2
    $\begingroup$ "I understand the partial products which do not share a common factor"... since you aren't working exclusively with integers, "factors" are now meaningless as everything is a factor of everything else (except zero). $\endgroup$
    – JMoravitz
    Apr 7 '20 at 13:11
17
$\begingroup$

You can separate the six products into three pairs with each pair having different factors $$x_1\cdot x_2\quad \ x_3\cdot x_4\\ x_1\cdot x_3\quad \ x_2 \cdot x_4\\ x_1 \cdot x_4\quad \ x_2 \cdot x_3$$ When we multiply the partial products on each line, we should get the same result. The only two pairs that have the same product are $2 \cdot 6$ and $3 \cdot 4$, so the product of the last line must also be $12$. The sixth partial product is $$\frac {12}5$$ Now we can check that the solution works. By symmetry we can assign the first line $2 \cdot 6,$ the second $3 \cdot 4$ and the last $5 \cdot \frac {12}5$ but we cannot be sure of the order of the last.Then $\frac {x_3}{x_2}=\frac 32.$ If the last is $5 \cdot \frac {12}5$ then $\frac {x_4}{x_2}=\frac 52, \frac {x_1}{x_2}=\frac 54$. The product of them all is $12$, so we have $$\frac 32\cdot \frac 52 \cdot \frac 54 x_2^4=12\\x_2=\sqrt{\frac 85}\\ x_1=\frac 54\sqrt {\frac 85}\\x_3=\frac 32\sqrt{\frac 85}\\x_4=\frac 52 \sqrt{\frac 85}$$ If we switch the products in the last line, we get another solution $$x_2=\sqrt{\frac {10}3}\\ x_1=\sqrt {\frac 65}\\x_3=\sqrt{\frac{15}2}\\x_4=\sqrt{\frac{24}5}$$ We can permute the assignment of the variables at will.

$\endgroup$
1
  • $\begingroup$ Stunning wave of thought! Really astounding idea!!! $\endgroup$ Apr 7 '20 at 14:02
4
$\begingroup$

We have the linear system $$ \left( {\matrix{ 1 & 1 & 0 & 0 \cr 1 & 0 & 1 & 0 \cr 1 & 0 & 0 & 1 \cr 0 & 1 & 1 & 0 \cr 0 & 1 & 0 & 1 \cr 0 & 0 & 1 & 1 \cr } } \right)\left( {\matrix{ {\ln x_{\,1} } \cr {\ln x_{\,2} } \cr {\ln x_{\,3} } \cr {\ln x_{\,4} } \cr } } \right) = \left( {\matrix{ {\ln p_{\,1} } \cr {\ln p_{\,2} } \cr {\ln p_{\,3} } \cr {\ln p_{\,4} } \cr {\ln p_{\,5} } \cr {\ln p_{\,6} } \cr } } \right)\quad \Rightarrow \quad {\bf M}\;{\bf y} = {\bf p} $$ The rank of the coefficient matrix is $4$, and so the rank of the augmented matrix shall be not greater than $4$.
This pose a constraint on the values and on the order of the $p_k$.

We can have a better look to the situation if we perform a Gaussian elimination on the augmented matrix.
We do that by performing a LU decomposition of $\bf M$ and then right-multiplying by the inverse of the L component, and arrive to $$ \left( {\matrix{ 1 & 1 & 0 & 0 \cr 0 & { - 1} & 1 & 0 \cr 0 & 0 & { - 1} & 1 \cr 0 & 0 & 0 & 2 \cr 0 & 0 & 0 & 0 \cr 0 & 0 & 0 & 0 \cr } } \right)\left( {\matrix{ {\ln x_{\,1} } \cr {\ln x_{\,2} } \cr {\ln x_{\,3} } \cr {\ln x_{\,4} } \cr } } \right) = \left( {\matrix{ 1 & 0 & 0 & 0 & 0 & 0 \cr { - 1} & 1 & 0 & 0 & 0 & 0 \cr 0 & { - 1} & 1 & 0 & 0 & 0 \cr { - 1} & { - 1} & 2 & 1 & 0 & 0 \cr 0 & 1 & { - 1} & { - 1} & 1 & 0 \cr 1 & 0 & { - 1} & { - 1} & 0 & 1 \cr } } \right)\left( {\matrix{ {\ln p_{\,1} } \cr {\ln p_{\,2} } \cr {\ln p_{\,3} } \cr {\ln p_{\,4} } \cr {\ln p_{\,5} } \cr {\ln p_{\,6} } \cr } } \right) $$

It is evident that the system is solvable, and has a unique solution, if and only if the $p_k$'s satisfy the last two equations, i.e. if $$ \left\{ \matrix{ \ln p_{\,2} + \ln p_{\,5} = \ln p_{\,3} + \ln p_{\,4} \hfill \cr \ln p_{\,1} + \ln p_{\,6} = \ln p_{\,3} + \ln p_{\,4} \hfill \cr} \right.\quad \Rightarrow \quad p_{\,1} p_{\,6} = p_{\,2} p_{\,5} = p_{\,3} p_{\,4} $$ and this gives a general method to solve this type of problems.

With the data you give, the only way to have two couples with the same product is $2 \cdot 6 = 3 \cdot 4$, and therefore $$ 2 \cdot 6 = {{12} \over 5} \cdot 5 = 3 \cdot 4 $$ which means $$ {\bf p}^{\,T} = \left( {\ln 2,\ln \left( {{{12} \over 5}} \right),\ln 3,\ln 4,\ln 5,\ln 6} \right) $$ Then the upper four equations are readily solved to give $$ {\bf y} = {1 \over 2}\left( {\matrix{ {\ln {6 \over 5}} \cr {\ln {{10} \over 3}} \cr {\ln {{24} \over 5}} \cr {\ln {{15} \over 2}} \cr } } \right)\quad \Rightarrow \quad {\bf x} = \left( {\matrix{ {\sqrt {{6 \over 5}} } \cr {\sqrt {{{10} \over 3}} } \cr {\sqrt {{{24} \over 5}} } \cr {\sqrt {{{15} \over 2}} } \cr } } \right) $$ and in fact $$ \eqalign{ & \sqrt {{6 \over 5}} \sqrt {{{10} \over 3}} = 2\quad \sqrt {{6 \over 5}} \sqrt {{{24} \over 5}} = {{12} \over 5}\quad \sqrt {{6 \over 5}} \sqrt {{{15} \over 2}} = 3 \cr & \sqrt {{{10} \over 3}} \sqrt {{{24} \over 5}} = 4\quad \sqrt {{{10} \over 3}} \sqrt {{{15} \over 2}} = 5\quad \sqrt {{{24} \over 5}} \sqrt {{{15} \over 2}} = 6 \cr} $$

However, swapping the values of $p_1$ and $p_6$ we get a different quadruple as result $$ {\bf x} = \left( {\matrix{ {\sqrt {{{18} \over 5}} } \cr {\sqrt {10} } \cr {\sqrt {{8 \over 5}} } \cr {\sqrt {{5 \over 2}} } \cr } } \right) $$ which checks as well.

I did not check for the other allowed permutations of $\bf p$.

$\endgroup$
3
$\begingroup$

Suppose the four numbers are $0<a<b<c<d$ Note: we can assume they are distinct since no products are repeated.

Let the six products be $P_1≤P_2≤\cdots ≤P_6$.

It is easy to see that these must be of the form $$ab<ac<\{bc,ad\}<bd<cd$$

Where the order of the two terms in the middle is uncertain.

If we assume that $P_6>6$ then we get $$\frac {P_1\times P_6}{P_2}=P_5\implies \frac {2P_6}{3}=6\implies P_6=9$$

But there are no solutions for $(a,b,c,d)$ consistent with this (brute force).

Similarly we can not have $P_1<2$.

If we had $P_2$ as the missing term then we'd get $$\frac {2\times 6}{5}=P_2\implies \boxed {P_2=\frac {12}5}$$

That one works! Indeed we could have $$a=2\sqrt {\frac 25}\quad b=\sqrt {\frac 52}\quad c= 3\sqrt {\frac25}\quad d=\sqrt {10}$$

I did not try to analyze the other cases, though this would be no harder.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.