3
$\begingroup$

Also, $x$,$y$ and $z$ are positive.

I have figured out that it would be enough to prove that LHS is bigger than $(27-\text{LHS})/3$, then by substituting 27 with $ (x+y+z)^3 $, it's enough to prove that $$x^3+y^3+z^3+6xyz \geq x^2(3-x)+y^2(3-y)+z^2(3-z).$$

And I'm stuck. I feel like this has to be provable somehow with rearrangement, since the right hand side is conveniently ordered (if $x>y$, then $3-x<3-y$). But the equality only holds when one of the variables is $0$ and the other two are $1.5$, which isn't the case for rearrangement.

$\endgroup$
4
  • $\begingroup$ Let $f(x,y,z)$ be the function on the LHS. First show that the minimum of $f$ can be achieved when some number among $x,y,z$ equals 0. Secondly, set $z=0$ without loss of generality and show that the minimum of $f(x,3-x,0)$ where $x$ varies from 0 to 3, is exactly 27/4. You may do this in many ways, including the familiar method of calculus of a single variable. $\endgroup$
    – Dilemian
    Commented Jan 15, 2017 at 0:51
  • $\begingroup$ I did try taking the derivative of LHS in respect to x, it's 3x^2 + 6*yz, and in order for that to be 0, x and one of y,z should be 0, which would give the maximum of the function. Can you elaborate on how you would do it please, I have no knowledge of calculus with several variables and know very little calculus in general. $\endgroup$
    – TAJD
    Commented Jan 15, 2017 at 1:09
  • $\begingroup$ Nevermind I just figured out that since the derivative is positive x=0 at the minimum. Thanks. $\endgroup$
    – TAJD
    Commented Jan 15, 2017 at 1:16
  • $\begingroup$ You cannot just take derivative of LHS in respect to $x$ without using the restriction of $x+y+z$. I will provide an answer in a minute. $\endgroup$
    – S. Y
    Commented Jan 15, 2017 at 5:56

2 Answers 2

1
$\begingroup$

Without loss of generality, we assume $z\ge \frac{x+y+z}{3}=1$. We temporarily fix $z$ and check how to arrange $x$ and $y$ to achieve a minimum. Now $y=3-x-z$.

So $f(x)=x^3+(3-x-z)^3+z^3+6x(3-x-z)z$, and the derivative of $f$ is $$f'(x) = 3x^2-3(3-x-z)^2 + 6(3-x-z) z-6xz=x(-3*2*(z-3) -6z-6z) + (-3(z-3)^2+6(3-z)z) = x(-18z+18) + 3(3-z)(9z-9) $$ and $f''(x)=-18z+18 \le 0$, which means $f(x)$ assumes its minimum at the boundary of $[0, 3-z]$, then either $x=0$ or $y=0$. Without loss of generality, we assume $y=0$, and now we want to prove $x^3+z^3\ge 27/4$ given $x+z=3$ and $x\ge 0$ and $z\ge 0$, which is much easier than the original problem.

$\endgroup$
0
$\begingroup$

We need to prove that $$\sum\limits_{cyc}(4x^3+8xyz)\geq(x+y+z)^3$$ or $$\sum_{cyc}(x^3-x^2y-x^2z+xyz)+3xyz\geq0,$$ which follows from Schur.

Done!

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .