3
$\begingroup$

If $\log _b a\cdot\log_c a +\log _a b\cdot\log_c b+\log _a c\cdot\log_b c=3$ and $a,b,c$ are different positive real numbers not equal to 1, then find the value of $abc$.

I tried to simplify this by different methods like using the identity $\log_b a=\frac{1}{\log_a b}$, but I couldn't get anywhere. Addition, subraction can't be used. I am not able to figure out a way to simplify the LHS. It would be great if someone could help me to proceed with this problem.

$\endgroup$
1
  • 5
    $\begingroup$ Hint: $\log_a b = \frac{\ln b }{\ln a}$. Apply this to all terms, then simplify using log rules. $\endgroup$ Commented Dec 15, 2016 at 12:43

4 Answers 4

4
$\begingroup$

If you set $X=\log a$, $Y=\log b$ and $Z=\log c$ (any base you like), your condition becomes easily $$ \frac{X^3+Y^3+Z^3}{XYZ}=3 $$ Now you can use the identity $$ X^3+Y^3+Z^3=(X+Y+Z)^3-3(X+Y+Z)(XY+YZ+ZX)+3XYZ $$ to get that $$ (X+Y+Z)\bigl((X+Y+Z)^2-3(XY+YZ+ZX)\bigr)=0 $$ The second factor can be written $$ X^2+Y^2+Z^2-XY-YZ-ZX $$ and we should consider the quadratic form having as matrix $$ A=\begin{bmatrix} 1 & -1/2 & -1/2 \\ -1/2 & 1 & -1/2 \\ -1/2 & -1/2 & 1 \end{bmatrix} $$ Since $$ \det[1]=1>0, \qquad \det\begin{bmatrix}1 & -1/2 \\ -1/2 & 1\end{bmatrix}=3/4>0 \qquad \det A=0 $$ the quadratic form is positive semidefinite. Its null space contains the vector $[1\;1\;1]^T$, so we can conclude that your condition implies $$ X+Y+Z=0\qquad\text{or}\qquad X=Y=Z $$ Since by assumption $a$, $b$ and $c$ are pairwise distinct, we end with $X+Y+Z=0$.


Without quadratic forms, you can reason about $X^2+Y^2+Z^2-XY-YZ-ZX=0$ as follows. Since $Z\ne0$ by assumption, we can set $X=uZ$ and $Y=vZ$, so the equation becomes $$ u^2-uv+v^2-u-v+1=0 $$ and, solving with respect to $v$, $v^2-v(u+1)+u^2-u+1=0$, the discriminant is $$ (u+1)^2-4(u^2-u+1)=-3(u-1)^2 $$ so the equation has a solution only for $u=1$, which gives $v=1$. Therefore $X=Y=Z$.


Another “elementary” approach. Suppose $(X+Y+Z)^2=3(XY+YZ+ZX)$. Set $s=X+Y+Z$ and $p=XYZ$; then $X$, $Y$ and $Z$ are the roots of the equation $$ t^3-st^2+\frac{s^2}{3}t-p=0 $$ by Viète's formulas. We can complete the cube getting $$ \left(t-\frac{s}{3}\right)^3=p-\frac{s^3}{27} $$ which should have three real roots. This is impossible unless the roots are coincident, so $p=s^3/27$ and $X=Y=Z$.

$\endgroup$
7
  • $\begingroup$ Can't be understood If you don't know linear algebra, I suggest to replace it with a non-negative discriminant argument. $\endgroup$ Commented Dec 15, 2016 at 14:05
  • $\begingroup$ @i9Fn Agreed, but first of all I wanted to write a correct answer, which the others aren't. $\endgroup$ Commented Dec 15, 2016 at 14:10
  • $\begingroup$ I cannot express how much I appreciate this answer. I tried to write it myself, and realized I didn't know how. It definitely fills out the interesting (and more challenging) part of the original problem. $\endgroup$ Commented Dec 15, 2016 at 14:59
  • $\begingroup$ @JohnHughes Quadratic forms are a big fun. ;-) $\endgroup$ Commented Dec 15, 2016 at 15:39
  • $\begingroup$ Oh, yeah. Esp. over $\Bbb Z/ 2\Bbb Z$, where there's the Arf invariant. :) $\endgroup$ Commented Dec 15, 2016 at 15:42
2
$\begingroup$

We can write $\log_{b}a = \frac{\log a}{\log b}$ and similarly $\log_{c}a = \frac{\log a}{\log c}$. Expanding like this, we have, $$\log_ba\log_ca + \log_ab\log_cb + \log_ac\log_bc =3$$ $$\Rightarrow \frac{(\log a)^2}{\log b\log c} + \frac{(\log b)^2}{\log a\log c} + \frac{(\log c)^2}{\log a\log b} =3$$ $$\Rightarrow \frac{(\log a)^3 + (\log b)^3 + (\log c)^3}{\log a\log b\log c} =3$$ $$\Rightarrow (\log a)^3 + (\log b)^3 + (\log c)^3 = 3\log a\log b\log c...(1)$$ Let $\log a =x, \log b=y, \log c =z$. Then equation $(1)$ transforms to $$x^3+y^3+z^3 = 3xyz \Leftrightarrow x+y+z=0$$ Then, $$\log a+\log b+\log c=0$$ $$\Rightarrow \log abc =0$$ $$\Rightarrow abc =10^0=1$$ Hope it helps.

$\endgroup$
6
  • 4
    $\begingroup$ Also: the fact that $x+y+z = 0$ solves the equation doesn't mean that every solution has the property that $x +y+z = 0$. $\endgroup$ Commented Dec 15, 2016 at 13:02
  • $\begingroup$ That double-implication is not the result of transforming the previous equation as claimed; it's a separate result, and requires justification. $\endgroup$ Commented Dec 15, 2016 at 13:47
  • $\begingroup$ @JohnHughes Also see Roman83's answer below. $\endgroup$ Commented Dec 15, 2016 at 13:49
  • $\begingroup$ Right: Roman83 skips the same step that you do. That doesn't make it right. :) $\endgroup$ Commented Dec 15, 2016 at 13:51
  • 1
    $\begingroup$ Also note that your conclusion that $x^3 + y^3 + z^3 = 3xyz$ is equivalent to $x+y+z = 0$ is simply false: consider the case $x = y = z = 1$. $\endgroup$ Commented Dec 15, 2016 at 15:52
1
$\begingroup$

Use $\log_a b=\frac{\ln a}{\ln b}$ $$\log _b a\cdot\log_c a +\log _a b\cdot\log_c b+\log _a c\cdot\log_b c=3$$ $$\frac{\ln a}{\ln b}\cdot \frac{\ln a}{\ln c}+\frac{\ln b}{\ln a}\cdot\frac{\ln b}{\ln c}+\frac{\ln c}{\ln a}\cdot\frac{\ln c}{\ln b}=3$$ $$\ln^3a+\ln^3b+\ln^3c=3(\ln a\ln b\ln c)$$ But $x^3+y^3+z^3=3xyz \Leftrightarrow x+y+z=0$

Then $\ln a+\ln b+\ln c=0$ $$\ln abc=0$$ Then $$abc=1$$

$\endgroup$
1
  • $\begingroup$ Your conclusion that $x^3 + y^3 + z^3 = 3xyz$ is equivalent to $x+y+z = 0$ is simply false: consider the case $x = y = z = 1$. (This possibility, in the problem, is ruled out by the assumption that $a,b,c$ are distinct, an assumption that your solution never uses.) $\endgroup$ Commented Dec 15, 2016 at 15:53
0
$\begingroup$

Got it!
$\log _b a\cdot\log_c a +\log _a b\cdot\log_c b+\log _a c\cdot\log_b c=3$
$$\frac{{(\log a)}^2}{\log b\cdot\log c} +\frac{{(\log b)}^2}{\log a\cdot\log c} +\frac{{(\log c)}^2}{\log a\cdot\log b} =3$$

$$\frac{{(\log a)}^3+{(\log b)}^3+{(\log c)}^3}{\log a\log b\log c} =3$$

$${(\log a)}^3+{(\log b)}^3+{(\log c)}^3=3\log a\log b\log c$$

This means $ \log a+\log b+\log c=0 \Rightarrow \log{abc}=0$
Hence, $abc=1$

$\endgroup$
2
  • $\begingroup$ The jump to "this means that..." could use a little justification; otherwise, you're good. $\endgroup$ Commented Dec 15, 2016 at 13:03
  • $\begingroup$ See egreg's answer to fill in the justification for the "this means that". $\endgroup$ Commented Dec 15, 2016 at 15:00

You must log in to answer this question.

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.