An Easy Way And A Hard Way II

Algebra Level 4

9 a 2 + 9 a b a b + b c + 6 b 2 + 6 b c a c + b c + 4 c 2 + 4 a c a b + a c \large \frac{9a^{2}+9ab}{ab+bc}+\frac{6b^{2}+6bc}{ac+bc}+\frac{4c^{2}+4ac}{ab+ac}

Given that a , b , c a,b,c are positive real numbers, find the minimum value of the expression above.


The answer is 18.

This section requires Javascript.
You are seeing this because something didn't load right. We suggest you, (a) try refreshing the page, (b) enabling javascript if it is disabled on your browser and, finally, (c) loading the non-javascript version of this page . We're sorry about the hassle.

1 solution

Brandon Monsen
Nov 7, 2015

Let the expression = k =k . We can rewrite k k as:

9 a ( a + b ) b ( a + c ) + 6 b ( b + c ) c ( a + b ) + 4 c ( a + c ) a ( b + c ) = k \frac{9a(a+b)}{b(a+c)}+\frac{6b(b+c)}{c(a+b)}+\frac{4c(a+c)}{a(b+c)}=k

Since, a , b , c > 0 a,b,c>0 , any combinations of sums, product, and quotients of a , b , c a,b,c will be positive, so we can use the AM-GM inequality. This gets us:

k 3 9 ( a ) ( a + b ) ( 6 ) ( b ) ( b + c ) ( 4 ) ( c ) ( a + c ) b ( a + c ) ( c ) ( a + b ) ( a ) ( b + c ) 3 \frac{k}{3} \geq \sqrt[3]{\frac{9(a)(a+b)(6)(b)(b+c)(4)(c)(a+c)}{b(a+c)(c)(a+b)(a)(b+c)}} .

Everything under the cube root cancels except for 9 × 6 × 4 9 \times 6 \times 4 , so we end up with:

k 3 216 3 \frac{k}{3} \geq \sqrt[\large 3]{216}

Solving for k k gives:

k 18 \large k \geq 18

so our answer is

18 \large \boxed{18} .


I deleted the extension of the solution in order to prevent people from instantly getting the answer to my new problem, A Hard Way And A Hard Way

Can you show us that this minimum is actually attained?

Otto Bretscher - 5 years, 7 months ago

Log in to reply

Sure thing. I haven't tried to prove it, nor am I sure that I could, but these were the results that I got from Wolfram Alpha. Ill work on it by hand in the meantime and see if I get anywhere.

Brandon Monsen - 5 years, 7 months ago

Log in to reply

The results from Wolfram Alpha are not fully convincing, of course, since we must presume that they are rounded. I just wanted to point out that the problem is not that easy after all... finding the lower bound 18 is the easy part.

I did the math, out of curiosity... let's see whether our solutions match. I will let you go first since it is your problem ;)

Otto Bretscher - 5 years, 7 months ago

Log in to reply

@Otto Bretscher Sorry about the late reply, I just saw your response. I ended up short of a solution as of yet, but here is what I have so far:

deleted to make sure nobody gets the solution to my new problem for free

Brandon Monsen - 5 years, 7 months ago

Log in to reply

@Brandon Monsen Yes, you are on the right track, in fact, you are almost done. To make life a little easier, you can assume that a = 1 a=1 since ( a , b , c ) (a,b,c) can be scaled without changing the answer.

Otto Bretscher - 5 years, 7 months ago

Log in to reply

@Otto Bretscher deleted to make sure nobody gets the answer to my new problem for free

Brandon Monsen - 5 years, 7 months ago

Log in to reply

@Brandon Monsen Nicely done. You should post this in your solution as well.

Bonus : Find the exact ratio a : b : c a:b:c .

Pi Han Goh - 5 years, 7 months ago

@Brandon Monsen Yes, exactly, that is the idea!

Just a small point to clarify: When you say "we can set these two [equations] equal to each other", you are not actually showing that both quantities are zero.

Otto Bretscher - 5 years, 7 months ago

Log in to reply

@Otto Bretscher deleted to make sure nobody gets the answer to my new problem for free

Brandon Monsen - 5 years, 7 months ago

Log in to reply

@Brandon Monsen There is a logical flaw in your reasoning: You cannot show that p ( x ) = 0 p(x)=0 and q ( x ) = 0 q(x)=0 merely by showing that p ( x ) = q ( x ) p(x)=q(x) . Luckily, it's easy to fix.

Otto Bretscher - 5 years, 7 months ago

Log in to reply

@Otto Bretscher deleted to make sure nobody gets the answer to my new problem for free

Brandon Monsen - 5 years, 7 months ago

Log in to reply

@Brandon Monsen Exactly! I'm sorry for being so insistent, but this is a beautiful problem and the solution has to be complete and flawless!

Otto Bretscher - 5 years, 7 months ago

Log in to reply

@Otto Bretscher Don't worry about it! I had just learned about the am-gm inequality a few days ago, and if you hadn't called me out on it I would've thought that it worked no matter what, and never checked to see if that maximum or minimum was possible to obtain. Thank you! Ill update my solution when I finish my homework :)

Brandon Monsen - 5 years, 7 months ago

Do you know what is the "hard way"?

Arihant Samar - 5 years, 7 months ago

Log in to reply

The reason I used this name was because I posted a problem a bit before this one by the same name that seemed super complicated and in-depth but ended up having a very short and simple solution. To answer your question: No I do not. I'm sure you could use calculus and some partial derivatives to look at what happens to the expression when you change either a , b , a,b, or c c . You would want to set each of those partial derivatives equal to 0, and you would end up with a system of three equation and three unknowns, for which you might be able to solve, but that would be incredibly tedious. I guess hence the name :P

Brandon Monsen - 5 years, 7 months ago

Why we have to check for p=q=r

Ishan Das - 5 years, 7 months ago

Log in to reply

deleted to make sure nobody gets the answer to my new problem for free

Brandon Monsen - 5 years, 7 months ago

Brandon, is this the easy way or the hard one? I guess this is the hard one but when will you show your easy way again?

Debmeet Banerjee - 5 years, 7 months ago

Log in to reply

I'm sorry about the late reply! I didn't see your comment until just now. The easy way is simply using AM-GM to find where the minimum is. The hard way would be using partial derivatives to determine minimums. The reason that the solution looked so hard is because the AM-GM inequality only gives a lower bound, not a true minimum value. For the purpose of a numerical answer, only AM-GM is necessary. On an Olympiad however, or just for the sake of a good solution, you have to show that the minimum us actually obtained, as Otto pointed out.

Brandon Monsen - 5 years, 6 months ago

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...