Waste less time on Facebook — follow Brilliant.
×

Not true!

Let \(a,b,c\) be real numbers such that \(a+b+c=0\) and \(a^2+b^2+c^2=1.\)Prove that \(a^2b^2c^2\leq \frac{1}{54}.\)When does this equality hold true?

Note by Ayush Rai
1 year, 4 months ago

No vote yet
1 vote

Comments

Sort by:

Top Newest

Without losing generality we assume that \(ab\geq 0 \). From the 1st condition we have \(a+b=-c\) replace this into the 2nd condition \[a^2+b^2+(a+b)^2=1\] By C-S \(a^2+b^2\geq\frac{(a+b)^2}{2}\), so \(1\geq\frac{3}{2}(a+b)^2 \Rightarrow c^2\leq\frac{2}{3}\). Now we rewrite the 2nd condition \[2(a^2+b^2)+2ab=1\] We have this inequality for all reals \(a^2+b^2\geq 2ab\), therefore \(6ab\leq 1\Rightarrow a^2b^2\leq\frac{1}{36}\) \[\therefore a^2b^2c^2\leq\frac{1}{54}\] The equality holds when \((a,b,c)=\bigg(\frac{1}{\sqrt{6}},\frac{1}{\sqrt{6}},\frac{-2}{\sqrt{6}}\bigg);\bigg(\frac{-1}{\sqrt{6}},\frac{-1}{\sqrt{6}},\frac{2}{\sqrt{6}}\bigg)\)

Gurīdo Cuong - 1 year, 4 months ago

Log in to reply

Let \(\alpha=abc\).

Now, as \[0=(a+b+c)^2=a^2+b^2+c^2+2(ab+bc+ca)\\\implies 0=1+2(ab+bc+ca)\\\implies ab+bc+ca=-\frac{1}{2}\]

Therefore, \(a, b, c\) are the (real) roots of the following function:

\[f(x)=x^3-\frac{1}{2}x-\alpha\]

For the existence of three real roots, we must have: \[f(x_1)f(x_2) \leq 0\] where \(x_1\) and \(x_2\) are the points of extrema of the function \(f(x)\).

Note that \(f'(x_1)=f'(x_2)=0\).

Since \(f'(x)\) is a quadratic polynamial, it can be solved.

Solving, plugging it into the above inequality, we get: \[|\alpha| \leq \frac{1}{3\sqrt6} \\\implies \alpha^2 \leq \frac{1}{54} \]

With equality holding if and only if two of the real numbers \(a, b, c\) are equal.

Deeparaj Bhat - 1 year, 4 months ago

Log in to reply

Note: It's slightly more straightforward to just take the cubic discriminant condition for 3 real roots.

Calvin Lin Staff - 1 year, 4 months ago

Log in to reply

The problem was I didn't remeber the cubic discriminant formula then. So, I used calculus in the solution.

Deeparaj Bhat - 1 year, 4 months ago

Log in to reply

Comment deleted Jun 07, 2016

Log in to reply

Btw, equality holds only if two of the three numbers \(a, b, c\) are equal.

Deeparaj Bhat - 1 year, 4 months ago

Log in to reply

You can't use am-gm inequality as \(a, b\) and \(c\) aren't non negative (if they are, then the inequality to be proved holds trivially).

Since \(a, b, c\) are real, \(a^2b^2c^2 \geq 0\).

Deeparaj Bhat - 1 year, 4 months ago

Log in to reply

Right, actually, if the three are non-negative, then the three are equal to 0. But this can be proved by cases, for example, WLOG let's assume \(a\) and \(b\) are positive and \(c\) is negative, then \(-c\) is positive, and so on.

I got a question... if you post a problem as a note, why to give yourself the solution?

Mateo Matijasevick - 1 year, 4 months ago

Log in to reply

@Mateo Matijasevick because it is a proving problem

Ayush Rai - 1 year, 4 months ago

Log in to reply

ya you are right

Abhishek Alva - 1 year, 4 months ago

Log in to reply

@Abhishek Alva i know you did not undertand anything.Just interrupting for nothing.

Ayush Rai - 1 year, 4 months ago

Log in to reply

ok @Deeparaj Bhat

Ayush Rai - 1 year, 4 months ago

Log in to reply

×

Problem Loading...

Note Loading...

Set Loading...