let $x,y\ge 0$,show that $$4x^3+8y^3+15xy^2-27x-54y+54\ge 0$$

when $x=y=1$ is equality.

see: http://www.wolframalpha.com/input/?i=4x%5E3%2B8y%5E3%2B15xy%5E2-27x-54y%2B54

this inequality is creat by me.and maybe have some methods to prove it? Thank you


We can prove that in part of the region $x \ge 0, y \ge 0$, the inequality holds$.

Let $f(x,y) = 4x^3+8y^3+15xy^2-27x-54y+54$

If $y=x$ then $$f(x,x)=27(-1 + x)^2(2 + x)\ge 0$$.

Set $x=y+a$ with $a=\frac{3^{3/2}}{2}$, then for $y\ge 0$ we have: $$f(y+a,y)=54 + \frac{81\sqrt{3}}{2}y^2 + 27y^3\ge 0$$

Set $y=x+b$ with $b=\frac{3^{3/2}}{13^{1/2}}$, then for $x\ge 0$ we have: $$f(x,x+b)=\frac{27}{169}\left(338 - 54\sqrt{39} + 78\sqrt{39}x^2 + 169x^3\right)\ge 0$$

This is because $$ 338^2=114244>113724= (54\sqrt{39})^2$$.


Find the point/s where the gradient vanishes, check that these evaluate to $\geq 0$, and show these are minimas. In case of saddle points, show these evaluate to $>0$.

  • $\begingroup$ what? I want to see such AM-GM,sos,Cauchy-Schwarz ,Holder,and so on to prove it.. $\endgroup$ – math110 Jun 21 '14 at 15:41
  • $\begingroup$ @math110 Perhaps you want to state this in the question. Feel free to down vote :) $\endgroup$ – user76568 Jun 21 '14 at 15:45
  • $\begingroup$ Also check along $x=0$ and $y=0$, where it is $(2y-3)^2(2y+6)$ and $(2x-3)^2(x+3)+27$. $\endgroup$ – Empy2 Jun 21 '14 at 16:34
  • $\begingroup$ I get a saddle point at $(x,y)=(\sqrt{3/28},\sqrt{12/7})$. $\endgroup$ – Empy2 Jun 21 '14 at 16:56
  • $\begingroup$ @Michael Correct, but it evaluates to a positive number, and hence $(1,1)$ is the only minimum for $x,y \geq 0$ $\endgroup$ – user76568 Jun 21 '14 at 17:39

Change variables to $x=X+1,y=Y+1$, and I think you get $$4X^3+12X^2+8Y^3+39Y^2+15XY^2+30XY$$

which is positive if $X$ and $Y$ are positive - so if $x\geq 1,y\geq 1$.

If $X\geq 0$, then it is at least $12X^2+30XY+31Y^2$, which is always positive.

When $x<1,y<1$, then $F_x$ and $F_y$ are both negative, so $F(x,y)>F(1,1)$.

When $y>3/2$, $F_y(x,y)=24y^2+30xy-54>0$, so $F(x,y)>F(x,3/2)$

I still have $0<x<1,1<y<3/2$ to go.


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.