⎩ ⎨ ⎧ a + b + c = 0 a 2 + b 2 + c 2 = 1
Let a , b , c be real numbers satisfying the above system of equations. Find the maximum possible value of ( 9 a b c ) 2 .
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.
Instead of diving into cubics, I did it using quadratics.
a + b = − c
a 2 + b 2 = 1 − c 2 ⇒ a b = 2 2 c 2 − 1
Then a , b are the roots of the polynomial:
x 2 + c x + 2 2 c 2 − 1
Since a , b ∈ R , we must have:
c 2 − 4 c 2 + 2 ≥ 0
c 2 ≤ 3 2
c ≤ 3 2
Also,
a b = 2 2 c 2 − 1 ≤ 6 1
Therefore, a b c ≤ 6 1 × 3 2
⇒ 8 1 ( a b c ) 2 ≤ 2 3
Log in to reply
Interesting variation, Aditya.
I don't think anyone has done it my way (hopefully someone will correct me if I'm wrong) so I'll throw it in here as well ...
Claim: For a , b ≥ 0 and a 2 + b 2 + ( a + b ) 2 = 1 , the maximum value of a b ( a + b ) occurs when a = b .
Proof: Rewriting the above equation as a 2 + b 2 + a b = 1 / 2 , we note that the maximization of a b and the maximization of a + b both coincide with the minimization of a 2 + b 2 (so of course this will also maximize a b ( a + b ) ). But this is also equivalent to minimizing the ratio a b a 2 + b 2 = b a + a b , and it can be shown using single-variable calculus that this occurs when b a = 1 , as desired. QED
The answer 1.5 can be computed from there. (Note that a , b ≥ 0 is not in the original problem but can be assumed without loss of generality.)
This is 'THE' best approach!
P.S. I would only suggest one change, Aditya. Namely to begin by stating your assumptions on the signs of a , b , c .
More specifically, your ending works if you assume that a , b ≤ 0 and c ≥ 0 (an assumption which is allowable, WLOG).
But that aside, very nice solution!
and please have a look at my solution...
sir, how did you find D?
Log in to reply
The derivation of the discriminant of a cubic polynomial is a little hard, but for this form where there is no x 2 term it's easier. Remember that the discriminant of a monic polynomial is the square of all products of every difference between its roots. Have a look in wikipedia, there is more information about it.
What is the discriminant of a general 3degree eqn?
I followed the same method though.
If I told you the error I did, you'd smash your face. Nice solution.
@Harsh Shrivastava Isn't this problem overrated?
Log in to reply
I agree that it is overrated.
and have a look at my solution
Log in to reply
Your solution is not clear. I'd advise you to use LaTeX.
Log in to reply
Nah. No genius things here :P
No flaw. I was just saying that if I told you the Error I made, you'll smash your face, like, how can someone be so dumb.
But on a serious note. This can be a 160 point prob.
Log in to reply
@Mehul Arora – Ohk , BTW you are right the prob's overrated.
Consider a monic cubic polynomial with roots as a,b,c
P ( x ) = x 3 + S 1 x 2 + S 2 x + S 3
By Vieta’s Sums: a + b + c = − S 1 = 0 and a b + b c + c a = − 2 1 and a b c = − S 3
∴ P ( x ) = x 3 − 2 1 x + S 3
But P ′ ( x ) = 3 x 2 − 2 1
Critical points of this function are at: x = − + 6 1
for the function to have three REAL roots, the maxima and minima of the function MUST BE greater than and less than zero respectively
GRAPHICALLY: This can be easily understood by the fact that for the function to have three real roots, the
curve must cut the X-axis at three points, which is possible only if the curve BULGES OUTWARD of the axis,
which again, is only possible if the critical points lie alternately above and below the X axis for max. and min.
∴ P ( 6 1 ) > 0 ⟹ S 3 > 3 6 1 ⟹ a b c < − 3 6 1
∴ ( 9 a b c ) 2 < 1 . 5
I think P(-1/sqrt(6)) should be greater than 0 instead of P(1/sqrt(6))
here a+b+c=0 and a^2+b^2+c^2=1 that's why we get from (a+b+c)^2 formula that ab+bc+ca= -1/2 and then a(b+c)+bc= -1/2 => a(-a)+ X/a= -1/2 [ let abc= X ] then X= a^3- a/2 then differentiated by a parameter for getting maximum value of X (abc) then we get a= +- Sq. root (1/6) and put this value above equation and get (9X)^2= (9abc)^2= 3/2 or 1.5
Hmm nice work!
First let's semplify a bit the system of equation:
{ a + b + c = 0 a 2 + b 2 + c 2 = 1
{ c = − ( a + b ) a 2 + b 2 + [ − ( a + b ) ] 2 = 1
{ c = − ( a + b ) a 2 + b 2 + a 2 + 2 a b + b 2 = 1
{ c = − ( a + b ) 2 a 2 + 2 b 2 + 2 a b = 1
{ c = − ( a + b ) a 2 + b 2 + a b = 2 1
Second semplify what we are searching for:
( 9 a b c ) 2 = 3 4 a 2 b 2 c 2 = 3 4 a 2 b 2 [ − ( a + b ) ] 2 = 3 4 ( a b ) 2 ( a + b ) 2
Now we know from the equation a 2 + b 2 + a b = 2 1 that if a increase b mast decrease to keep true the equation. Also knowing that the product of a b is greater smoller is the difference between a and b , see if a = b (a-b=0, the smollest possible) is a solution and, if it is, what is the result. So be x = a = b :
x 2 + x 2 + x × x = 2 1
3 x 2 = 2 1
x 2 = 6 1
And finally find (9abc)^{2}:
3 4 ( a b ) 2 ( a + b ) 2 = 3 4 ( x × x ) 2 ( 2 x ) 2 = 3 4 ( x 2 ) 2 4 x 2 = 3 4 × ( 6 1 ) 2 × 6 4 = 2 3 = 1 . 5
So the solution is 1 . 5
y + z = − x
y 2 + z 2 = 1 − x 2
apply CS inequality we get
root(2/3) >= x
so let x = root(2/3)
then y + z = -root(2/3)
and
y 2 + z 2 = 1 / 3
so y = z = -root2/2.root3
solving gives answer 1.5
i applied CS inequality in this way :
2 ( y 2 + z 2 ) > = ( y + z ) 2
sorry, i suppose x = a, y = b and z = c... Any feedback? Harsh Srivastav
I initially tried to solve this problem using a Cauchy-Schwarz inequality. There must be an error in my approach, because it does not yield the correct solution. I was wondering if someone could take a look at my approach and identify my error. I am rather new to CS inequalities, so I am not always sure when they are applicable or what kind of assumptions need to be made.
I obtained the following identity using a similar approach to others: a b + a c + b c = − 2 1
I then obtained a new identity by squaring both sides of this equation:
a 2 b 2 + a 2 c 2 + b 2 c 2 + 2 a 2 b c + 2 a b 2 c + 2 a b c 2 = 4 1
a 2 b 2 + a 2 c 2 + b 2 c 2 + 2 a b c ( a + b + c ) = 4 1
a 2 b 2 + a 2 c 2 + b 2 c 2 + 2 a b c ( 0 ) = 4 1
a 2 b 2 + a 2 c 2 + b 2 c 2 = 4 1
I then set up the Cauchy-Schwarz inequality like so:
( a 2 b 2 + a 2 c 2 + b 2 c 2 ) ( c 2 + b 2 + a 2 ) ≥ ( a b c + a b c + a b c ) 2
( 4 1 ) ( 1 ) ≥ ( 3 a b c ) 2
4 1 ≥ ( 3 a b c ) 2
Multiplying both sides of the inequality by 9 yields:
4 9 ≥ ( 9 a b c ) 2
This implies that the maximum value of ( 9 a b c ) 2 is 4 9 , which is not true.
Ah... Just a common mistake when applying classical inequalities.
You have already assumed that at the maximum value, a 2 = b 2 = c 2 must be true. Which is actually false, because the first equation a + b + c = 0 can't be fulfilled.
In fact, the maximum value occurs at ( a 2 , b 2 , c 2 ) = ( 2 / 3 , 1 / 6 , 1 / 6 ) or any other permutation.
The common myth when solving classical inequalities is to assume that a = b = c . See relevant wiki: Inequalities with strange equality conditions .
From a²+b²+c²=1 We can replace a²+b²+c²= sin²(x)×(cos²(y)+sin²(y))+cos²(y) i.e. a=sin(x)sin(y),b=cos(x)sin(y),c=cos(y) So because a+b+c=0 we have sin(y)(sin(x)+cos(x))+cos(y)=0 I.e. sin(x)+cos(x)=-cos(y)/sin(y) So squaring that we get 1+sin(2x)=(1-sin²(y))/sin²(y) so sin(2x)=1/sin²(y)-2=(1-2sin²(y))/sin²(y) So 81a²b²c²=81sin²(x)cos²(x)sin⁴(y)cos²(y) =81/4×sin²(2x)sin⁴(y)(1-sin²(y))= 81/4×(1-2sin²(y))²(1-sin²(y)) Put sin²(y)=p, 0<=p<=1 So we need to find max of f(p)=81/4×(1-2p)²(1-p) i.e. f(p)=81/4×(-4p³+8p²-5p+1) So 0=f'(p)=81/4(-12p²+16p-5) p=1/2 or p=5/6 So we choose p=5/6 because f(0.5)=0 So max is 81/4×(1-10/6)²(1-5/6)= 81/4×4/9×1/6=3/2=1.5
My solution is quite intuitive though. Please correct me if I make some mistakes.
First notice that if one of them is 0 then the product is 0. If only one of them is negative and the rest are positive then the product is negative. Similarly, if only one of them is positive and the rest are negative then the product is positive.
Therefore, WLOG assume a,b < 0 and c > 0 c = − a − b c 2 = ( a + b ) 2 Thus, a 2 + b 2 + c 2 = 2 ( a 2 + b 2 + a b ) = 1 So, now we are maximizing 3 4 a 2 b 2 ( a 2 + b 2 + 2 a b ) = 3 4 a 2 b 2 ( a 2 + b 2 + a b + a b ) = 3 4 a 2 b 2 ( 1 / 2 + a b )
Let X = ab >0, so we are maximizing x 2 ( 1 / 2 + x ) which is an increasing function on X. That is, if we can maximize ab we will have the answer to this. By AM-GM inequality (note that AM-GM assume all values are positive but in our case they are all negative anyways, so we may assume a' = |a| >0 and everything work just fine) 1 / 6 = ( a 2 + b 2 + a b ) / 3 ≥ a b Therefore, the maximum occurs at a=b (x = 1/6) Thus
( 9 a b c ) 2 ≤ 3 4 ( 1 / 6 ) 2 ( 1 / 2 + 1 / 6 ) = 3 / 2
EDIT: I forgot the rule of thumb :D ... have to make sure those values exist. This is trivial: by AM-GM we know that a=b makes this work. That is |a|= |b| = 1/sqrt(6) (ie., a =b = -1/sqrt(6), c = 2/sqrt(6))
Lets proceed by trigonometry as all a,b,c have upper and lower limits. Solving for 'a' in terms of 'b' after eliminating 'c' using these two equations we get -sqrt(2/3)<= b sqrt(2/3). ( just make discriminant so obtained positive as 'a' is real.) so let; b=sqrt(2/3)(sin x). putitng this b back in 'a' and 'c' equation gives .................. 81aabbcc = (3/2)(sin x)(sin x) { (sin x)(sin x) - 3(cos x)(cos x) }. Clearly this maximum when sin x =1. which gives the maximum value as.......................1.5.
It isn't stupid, it shows that 1.5 is one value that occurs, namely when a=b. So then you could consider how to prove that it is the largest value that occurs.
Log in to reply
The a*b is maximized, when the a equals b. In terms, that we have a constant c, and the sum of a and b is a constant too. And a and b have the same sign.
Log in to reply
The a*b is maximized, when the a equals b.
Yes. That's one of the things I showed in my solution.
Problem Loading...
Note Loading...
Set Loading...
First we are going to find a b + a c + b c with the identity ( a + b + c ) 2 = a 2 + b 2 + c 2 + 2 ( a b + a c + b c ) :
a b + a c + b c = − 2 1
Then consider a monic polynomial P ( x ) with roots a , b and c . Then:
P ( x ) = ( x − a ) ( x − b ) ( x − c ) = x 3 − ( a + b + c ) x 2 + ( a b + a c + b c ) x − a b c
Substituting the known values:
P ( x ) = x 3 − 2 1 x − a b c
Now, for a , b and c to be real the discriminant of P must be greater than or equal to zero. Also, the discriminant of P ( x ) = x 3 + p x + q is just Δ = − 4 p 3 − 2 7 q 2 , so:
Δ = − 4 ( − 2 1 ) 3 − 2 7 ( − a b c ) 2 ≥ 0
4 ( − 8 1 ) + 2 7 ( a b c ) 2 ≤ 0
2 7 ( a b c ) 2 ≤ 2 1
( 9 a b c ) 2 ≤ 2 3