Let $a$, $b$ and $c$ be the sides of a triangle. Prove that $$a^\text{2}(b+c-a) + b^\text{2}(c+a-b) + c^\text{2}(a+b-c) \leqslant 3abc$$


I am very new and novice at this problem. I did a little try but couldn't succeed because I was unable to substitute the left term of the inequality into formula. I know a formula that

$a^\text{3} + b^\text{3} + c^\text{3} - 3abc$ = $(a+b+c)(a^\text{2} + b^\text{2} + c^\text{2} - ab -bc - ca)$. But how to use this formula in that case isn't known to me. And how to show the relation of the both side and when they will become equal?

A small help will be enough for me. Thanks in advance.

  • 3
    $\begingroup$ This is true for any non-negative numbers $a,b,c$ which do not have to side of triangles. Look up this for reference. $\endgroup$ – dezdichado Feb 18 at 17:21
  • $\begingroup$ @dezdichado I didn't know that. I obtained two benefit at the same time. Many gratefulness to you for the reference. $\endgroup$ – Anirban Niloy Feb 18 at 17:24
  • 1
    $\begingroup$ In the limiting case $a+b=c$ this becomes $2\leq3.$ I wonder if that's supposed to be a $2$ on the right-hand side. $\endgroup$ – saulspatz Feb 18 at 17:25
  • 1
    $\begingroup$ If I haven't made a mistake, when a,b,c are the sides of a triangle it can be shown that $a^2(b+c-a)+b^2(c+a-b)+c^2(a+b-c)>2abc$ The two sides are equal when $a+b=c$ and I differentiated the difference w.r.t $c$ assuming $a+b<c.$ $\endgroup$ – saulspatz Feb 18 at 17:47
  • 1
    $\begingroup$ @AngelaRichardson just rearrange the inequality and use Schur for when $n = 1.$ $\endgroup$ – dezdichado Feb 18 at 19:36

It's true for all non-negatives $a$, $b$ and $c$.

Indeed, since our inequality is symmetric, we can assume that $a\geq b\geq c$ and we obtain: $$3abc-\sum_{cyc}a^2(b+c-a)=\sum_{cyc}(a^3-a^2b-a^2c+abc)=\sum_{cyc}a(a-b)(a-c)\geq$$ $$\geq a(a-b)(a-c)+b(b-a)(b-c)=(a-b)^2(a+b-c)\geq0.$$


Using the Ravi substitution $$a=y+z,b=x+z,c=x+y$$ we have to prove that $$x^2y+x^2z+xy^2+xz^2+y^2z+yz^2-6xyz\geq 0$$ But this is AM-GM:

$$x^2y+x^2z+xy^2+xz^2+y^2z+yz^2\geq 6\sqrt[6]{x^6y^6z^6}=6xyz$$

  • $\begingroup$ @Sonnhard I don't know about the AM-GM. Would you please explain? Then it will be easy for me to clearify the point. $\endgroup$ – Anirban Niloy Feb 18 at 17:30
  • 1
    $\begingroup$ One only needs two variable AM/GM: $x^2y+yz^2\ge 2xyz$ etc. $\endgroup$ – Lord Shark the Unknown Feb 18 at 17:49
  • $\begingroup$ I use that $$\frac{a+b+c+d+e+f}{6}\geq \sqrt[6]{abcdef}$$ for $$a,b,c,d,e,f\geq 0$$ $\endgroup$ – Dr. Sonnhard Graubner Feb 18 at 18:10

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.