JEE Inequalities Starter

Algebra Level 3

For a , b , c > 0 a,b,c>0 and a + b + c = 6 a+b+c=6 . Find the minimum value of

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 \large \left (a+\frac{1}{b} \right )^{2}+ \left (b+\frac{1}{c} \right )^{2}+\left (c+\frac{1}{a} \right )^{2}


The answer is 18.75.

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.

8 solutions

Prasun Biswas
Feb 26, 2015

We first use Titu's Lemma on the given expression that yields us the following inequality:

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 ( a + b + c + 1 a + 1 b + 1 c ) 2 1 + 1 + 1 = ( 6 + 1 a + 1 b + 1 c ) 2 3 \left(a+\frac{1}{b}\right)^2+\left(b+\frac{1}{c}\right)^2+\left(c+\frac{1}{a}\right)^2 \geq \frac{\left(a+b+c+\frac{1}{a}+\frac{1}{b}+\frac{1}{c}\right)^2}{1+1+1}=\frac{\left(6+\frac{1}{a}+\frac{1}{b}+\frac{1}{c}\right)^2}{3}

Now, to minimize the RHS of this inequality so as to achieve the tightest lower bound, we need to minimize ( 1 a + 1 b + 1 c ) \left(\frac{1}{a}+\frac{1}{b}+\frac{1}{c}\right) . Now, we'll use the AM-HM inequality on the positive reals a , b , c a,b,c to obtain:

a + b + c 3 3 1 a + 1 b + 1 c 1 a + 1 b + 1 c 3 2 \frac{a+b+c}{3}\geq \frac{3}{\frac{1}{a}+\frac{1}{b}+\frac{1}{c}}\\ \implies \frac{1}{a}+\frac{1}{b}+\frac{1}{c}\geq ~\frac{3}{2}

and equality occurs at a = b = c = 2 a=b=c=2 .


So, using the minimum value obtained, we can evaluate the original inequality as follows:

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 75 4 = 18.75 \left(a+\frac{1}{b}\right)^2+\left(b+\frac{1}{c}\right)^2+\left(c+\frac{1}{a}\right)^2 \geq \frac{75}{4}=18.75

Now, we need to check the equality case to verify that this is indeed the minimum value of the expression. Using the equality case for Titu's Lemma, we can verify that equality occurs when,

( a + 1 b ) = ( b + 1 c ) = ( c + 1 a ) = 5 2 \left(a+\frac{1}{b}\right)=\left(b+\frac{1}{c}\right)=\left(c+\frac{1}{a}\right)=\frac{5}{2}

We can see that this equality case also occurs at a = b = c = 2 a=b=c=2 which was the same for the AM-HM inequality. Thus, the equality cases are matched and the minimum value of the original inequality is confirmed to be 18.75 \boxed{18.75}

Classic JEE style. a = b = c = 2 a=b=c=2

Raghav Vaidyanathan - 6 years, 3 months ago

Log in to reply

Actually, it's not necessary that the minimum will always occur when the variables are equal. That works for most cases where the function to be optimized is symmetric and you are given a symmetric constraint too but doesn't always guarantee minima. You need to present the full proof to conclude your solution.

You should check out this discussion.

Prasun Biswas - 6 years, 3 months ago

Log in to reply

I know that, I just commented on this situation. I substituted with full knowledge that I can go wrong... But under the circumstances, it was a very safe bet. Thanks.

Raghav Vaidyanathan - 6 years, 3 months ago

hi. i applied the AM>=GM method but didnt get the answer. please explain why. thanks :)

Lipsa Kar - 5 years, 7 months ago

Log in to reply

I cannot explain unless you show me your work. Just saying you applied AM>=GM doesn't help me in understanding where you went wrong. You can post your work as a reply to this comment and I'll try to point out the error in your work (if there is one).

Prasun Biswas - 5 years, 7 months ago

Log in to reply

i opened the brackets. so, i got this: a^2+1/b^2+2a/b+b^2+1/c^2+2b/c+c^2+1/a^2+2c/a. let this be equal to k. k has 9 terms in it. k/9>=(2^3)^(1/9) (AM>=GM) so, we get k=9x2^(1/3)=11.39. please clarify this. thanks

Lipsa Kar - 5 years, 7 months ago

Log in to reply

@Lipsa Kar The problem with this approach is that you're not considering the given equality constraint a + b + c = 6 a+b+c=6 . This approach could've been applicable if we didn't have the additional equality constraint (i.e., if we had a , b , c > 0 a,b,c\gt 0 with no further restrictions).


Interesting thing to note here is that even if we didn't have the equality constraint in the problem, the conclusion you get by that approach is flawed in the sense that, in that case, there would've been no minimum since there are no a , b , c > 0 a,b,c\gt 0 that would satisfy the equality condition of AM-GM, which would be,

a 2 = 1 b 2 = 2 a b = b 2 = 1 c 2 = 2 b c = c 2 = 1 a 2 = 2 c a a^2=\frac 1{b^2}=2\frac ab = b^2=\frac 1{c^2}=2\frac bc=c^2=\frac 1{a^2}=2\frac ca

It's easy to note that there are no solutions a , b , c > 0 a,b,c\gt 0 to this because the 1st, 4th and 7th equalities suggest that a = b = c a=b=c . Let that equal value (if it exists) be k k . You then get, using the above equality that k 2 = 2 k = 2 k^2=2\implies k=\sqrt 2 . Now, just plug this into the above equality case and you'll get a contradiction (specifically that 2 = 1 2 2=\frac 12 which is false). Hence, there wouldn't have been a minimum without the additional equality constraint and what you would've got is a strict inequality.

@Nihar Mahajan , you might find this example interesting, this is yet another case where not checking for the equality case can lead to incorrect conclusions.

Prasun Biswas - 5 years, 7 months ago

Please see my solution too!

Nihar Mahajan - 6 years, 3 months ago

Log in to reply

You've left out quite some parts. You just cannot substitute inequations without providing proper justification. Also, you haven't shown the equality case to verify whether the obtained minimum occurs or not within the domain and range of R \mathbb{R} . You need to show it in your solution, otherwise the proof is incomplete.

Prasun Biswas - 6 years, 3 months ago

Log in to reply

Root mean square inequality is applicable for all real numbers.Can you explain why there is need to verify the minimum value?The question asks for the minimum value and it is not important to show that minimum value occurs only when equality is obtained. Your solution proves that a = b = c = 2 a =b = c = 2 gives the minimum value.But there is no need to prove this because we only want the minimum value , which is obtained in my solution.Also , Can you tell me the reason why we cannot substitute the equations?

Nihar Mahajan - 6 years, 3 months ago

Log in to reply

@Nihar Mahajan You can substitute, but you need to provide the mathematical explanation for it. See what I did in my solution before making the substitutions! Also, the substitutions are only valid when the equality cases for both the inequations are the same.

Prasun Biswas - 6 years, 3 months ago

@Nihar Mahajan Consider the following expression: Z = ( x + 1 x ) , x > 0 Z=\left(x+\frac{1}{x}\right)~,x\gt 0 . We all know that the minimum value of Z Z is 2 2 , right? Now, let's complete the square in two different ways:

Z = ( x + 1 x ) 2 2 = ( x 1 x ) 2 + 2 Z=\left(\sqrt{x}+\frac{1}{\sqrt{x}}\right)^2-2=\left(\sqrt{x}-\frac{1}{\sqrt{x}}\right)^2+2

From the trivial inequality, one may instantly conclude that the minimum of Z Z is ( 2 ) (-2) , if one uses the first case. But that's wrong! So, how do we disprove the first case? The ambiguity is checked by verifying the equality case that shows the first case has minimum when x = ( 1 ) 0 x=(-1)\not\gt 0 , and for the second case, it occurs at x = 1 > 0 x=1\gt 0 . So, you see, it's very essential to verify the equality case since you need to show that the equality indeed occurs at a value of x x that is within the given domain of the problem. There are times when equality doesn't occur in a given inequality and then we get a strict inequality (without extrema).

Prasun Biswas - 6 years, 3 months ago

Log in to reply

@Prasun Biswas Thanks a lot!

Nihar Mahajan - 6 years, 3 months ago

@Prasun Biswas What to do in cases when the equality conditions don't match? Or even when they don't verify? Or when we only get a continuous lower/upper bound but not a minimum value?

Gaurav Manwani - 3 years, 11 months ago

After we open the brackets, we see cyclic order, the equation is: a^2 + 1/a^2 + b^2 +1/b^2 + c^2 + 1/c^2 + 2a/b + 2b/c + 2c/a

Now if we apply AM-GM, we get min. value to be 9*(9th root of 8)

But if we further split the 2a/b, 2b/c and 2c/a into individual a/b and so on, and then apply AM-GM we get ans as 12. In both the methods, we are not violating the terms that numbers need to be positive, every term is positive, so where is the error?

I know for a fact that I am mistaken somewhere as neither of my answers are correct, and is that wasn't enough, I got two different answers for same question LOL. But on a serious note, it will help a lot if you / anyone reading this could point out where have done the error.

Mrunang Kothari - 6 months, 3 weeks ago
Nihar Mahajan
Feb 27, 2015

We use the inequality R . M . S A . M R.M.S \geq A.M .

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 3 a + b + c + 1 a + 1 b + 1 c 3 \sqrt{\dfrac{\left(a + \frac{1}{b}\right)^2 + \left(b + \frac{1}{c}\right)^2 + \left(c + \frac{1}{a}\right)^2}{3}} \geq \dfrac{a + b + c + \dfrac{1}{a}+ \dfrac{1}{b}+ \dfrac{1}{c}}{3}

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 ( a + b + c + 1 a + 1 b + 1 c ) 2 3 ( 1 ) \Rightarrow\left(a + \frac{1}{b}\right)^2 + \left(b + \frac{1}{c}\right)^2 + \left(c + \frac{1}{a}\right)^2 \geq \dfrac{\left(a + b + c + \dfrac{1}{a}+ \dfrac{1}{b}+ \dfrac{1}{c}\right)^2}{3}\dots (1)

We also know that :

( a + b + c ) ( 1 a + 1 b + 1 c ) 3 2 (a + b + c)\left(\dfrac{1}{a}+ \dfrac{1}{b}+ \dfrac{1}{c}\right) \geq 3^2

( 6 ) ( 1 a + 1 b + 1 c ) 9 \Rightarrow(6)\left(\dfrac{1}{a}+ \dfrac{1}{b}+ \dfrac{1}{c}\right) \geq 9

( 1 a + 1 b + 1 c ) 9 6 \Rightarrow\left(\dfrac{1}{a}+ \dfrac{1}{b}+ \dfrac{1}{c}\right) \geq \dfrac{9}{6}

( 1 a + 1 b + 1 c ) 3 2 ( 2 ) \Rightarrow\left(\dfrac{1}{a}+ \dfrac{1}{b}+ \dfrac{1}{c}\right) \geq \dfrac{3}{2} \dots (2)

Substituting ( 2 ) (2) in ( 1 ) (1) ,

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 ( 6 + 3 2 ) 2 3 \Rightarrow\left(a + \frac{1}{b}\right)^2 + \left(b + \frac{1}{c}\right)^2 + \left(c + \frac{1}{a}\right)^2 \geq \dfrac{\left(6 + \dfrac{3}{2}\right)^2}{3}

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 ( 15 2 ) 2 3 \Rightarrow\left(a + \frac{1}{b}\right)^2 + \left(b + \frac{1}{c}\right)^2 + \left(c + \frac{1}{a}\right)^2 \geq \dfrac{\left(\dfrac{15}{2}\right)^2}{3}

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 ( 225 4 ) 3 \Rightarrow\left(a + \frac{1}{b}\right)^2 + \left(b + \frac{1}{c}\right)^2 + \left(c + \frac{1}{a}\right)^2 \geq \dfrac{\left(\dfrac{225}{4}\right)}{3}

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 225 12 \Rightarrow\left(a + \frac{1}{b}\right)^2 + \left(b + \frac{1}{c}\right)^2 + \left(c + \frac{1}{a}\right)^2 \geq \dfrac{225}{12}

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 18.75 \Rightarrow\left(a + \frac{1}{b}\right)^2 + \left(b + \frac{1}{c}\right)^2 + \left(c + \frac{1}{a}\right)^2 \geq \huge\boxed{18.75}

what inequality are you invoking to say that ( a + b + c ) ( ( 1 / a + 1 / b + 1 / c ) 3 2 (a+b+c)((1/a+1/b+1/c) \geq 3^2 ?

Mahoma Iza V - 4 years ago

Log in to reply

Cauchy-Schwarz:

wq y - 2 years, 12 months ago

2 inequalities are being invoked. 1) a+ b+c/3 is grter than or = to (abc)^1/3 (AM is grter than or = to GM) This simplifies to a+b+c greater than or = to 3(abc)^1/3 2) ( 1/a + 1/b+1/c)/3 greater than or equal to (1/a×1/b×1/c)^1/3, which again simplifies to 1/a+1/b +1/c greater or equal to 3 (........) Multiply these 2 ineqalities to get the result.

N L - 3 years, 11 months ago

Same solution

Hans Gabriel Daduya - 3 years, 7 months ago

Hey but why can't we just apply am gm after opening bracket

Atharva Verma - 1 year ago

Maybe I'm being dumb, but why is (a+b+c)(1/a+1/b+1/c)>=9?

Verda Glori - 2 months ago

Without titu's lemma= If we expand the original expression we get:

a 2 + 2 a b + 1 b 2 + b 2 + 2 b c + 1 c 2 + c 2 + 2 c a + 1 a 2 a^2+2\frac{a}{b}+\frac{1}{b^2}+b^2+2\frac{b}{c}+\frac{1}{c^2}+c^2+2\frac{c}{a}+\frac{1}{a^2}

reordering:

2 ( a b + b c + c a ) + ( a 2 + b 2 + c 2 ) + ( 1 a 2 + 1 b 2 + 1 c 2 ) 2*(\frac{a}{b}+\frac{b}{c}+\frac{c}{a})+(a^2+b^2+c^2)+(\frac{1}{a^2}+\frac{1}{b^2}+\frac{1}{c^2})

minimizing each of the three previous terms we get (using AM>=GM inequality):

2 ( 3 ) + 3 x + 3 1 x 2*(3)+3x+3\frac{1}{x} (equation z)

x = ( a b c ) 2 / 3 x=(abc)^{2/3}

using the identity given in the problem we have:

( a + b + c ) 2 = 6 2 (a+b+c)^2=6^2

a 2 + b 2 + c 2 = 36 2 ( a b + b c + a c ) a^2+b^2+c^2=36-2(ab+bc+ac)

if we try to minimize this we get:

36 2 3 x 36-2*3x

so now for the second term we have an easy equation:

36 6 x = 3 x ; x = 4 36-6x=3x; x=4

finally we return to equation z and we get:

6 + 3 4 + 3 ( 1 / 4 ) = 18.75 6+3*4+3*(1/4)=18.75

NOTE: First time using LaTeX :O

but why 36-6x is equal to 3x? It can be greater or less than it , isn't?

raghdah mansour - 9 months, 3 weeks ago

Because of symmetry, it is likely that a=b=c is some minimum. With a=b=c=2, the Exp. =18.75. We try a=1, b=2, c=3 and we get Exp. >18.75. So 18.75 \boxed{\color{#D61F06}{ \Large18.75 } } is the minimum.

First of all, it is likely that a = b = c a=b=c results in a minimum, but doesn't guarantee minimum value! Secondly, checking just a single case to see if it exceeds the suspected minima doesn't conclude that it is the minimum. You cannot conclude anything without a rigorous proof.

Prasun Biswas - 6 years, 3 months ago

Log in to reply

With a=b=c, it may be maximum or minimum. With the check , it shows to be minimum. Can any one please help me by explaining the condition when this works? And more about cyclic and symmetric. Thanks.

Niranjan Khanderia - 6 years, 3 months ago

Log in to reply

You should check out this discussion.

Prasun Biswas - 6 years, 3 months ago

Log in to reply

@Prasun Biswas Thank you for pointing to the article.

Niranjan Khanderia - 6 years, 3 months ago
Brian Albright
Mar 8, 2017

We use the RMS-AM followed by the AM-HM inequalities:

( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 3 ( a + 1 b ) + ( b + 1 c ) + ( c + 1 a ) 3 by RMS-AM = a + b + c + 1 a + 1 b + 1 c 3 = 6 3 + 1 a + 1 b + 1 c 3 2 + 3 a + b + c by AM-HM = 2 + 3 6 = 5 2 \begin{aligned} \sqrt {\frac{{{{\left( {a + \frac{1}{b}} \right)}^2} + {{\left( {b + \frac{1}{c}} \right)}^2} + {{\left( {c + \frac{1}{a}} \right)}^2}}}{3}} & \ge \frac{{\left( {a + \frac{1}{b}} \right) + \left( {b + \frac{1}{c}} \right) + \left( {c + \frac{1}{a}} \right)}}{3} \;\;\; \mbox{by RMS-AM}\\ & = \frac{{a + b + c + \frac{1}{a} + \frac{1}{b} + \frac{1}{c}}}{3}\\ & = \frac{6}{3} + \frac{{\frac{1}{a} + \frac{1}{b} + \frac{1}{c}}}{3}\\ & \ge 2 + \frac{3}{{a + b + c}} \;\;\; \mbox{by AM-HM}\\ & = 2 + \frac{3}{6} = \frac{5}{2} \end{aligned}

Squaring both sides and multiplying by 3 yields ( a + 1 b ) 2 + ( b + 1 c ) 2 + ( c + 1 a ) 2 75 4 = 18.75 {\left( {a + \frac{1}{b}} \right)^2} + {\left( {b + \frac{1}{c}} \right)^2} + {\left( {c + \frac{1}{a}} \right)^2} \ge \frac{{75}}{4} = 18.75

Magnas Bera
Jun 19, 2019

Use TITU'S LEMMA OR CAUCHY SCWARZ

Arijit Sinha
Jun 24, 2018

Use Cauchy-Schwarz inequallity followed by A.M. - H.M. inequality.

Emmanuel Torres
Feb 8, 2017

Just assume 2 for all of them and do the math. lol

0 pending reports

×

Problem Loading...

Note Loading...

Set Loading...