Let C 1 be the maximum value such that
x 2 + y 2 + 1 ≥ C 1 ( x + y )
for all real x and y . Similarly, let C 2 be the maximum value such that
x 2 + y 2 + x y + 1 ≥ C 2 ( x + y )
holds true for all real x and y . If C 1 C 2 can be expressed as n , find n .
This is really a proof problem in disguise. It's a very beautiful problem, and I hope you post a proof rather than just a brute-force solution explaining your answer.
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.
For the first inequality, observe that ( x − 2 2 ) 2 + ( y − 2 2 ) 2 ≥ 0 gives us
x 2 + y 2 + 1 ≥ 2 ( x + y ) .
That was easy, but doesn't afford an easy way of dealing with that x y term in the second inequality, so let's consider a different approach that allows us to relate it to x + y .
We have x 2 + y 2 = 2 1 ( x + y ) 2 + 2 1 ( x − y ) 2 . Hence, the first inequality is equivalent to
2 1 ( x + y ) 2 + 2 1 ( x − y ) 2 + 1 ≥ C 1 ( x + y ) .
By AM-GM and the trivial inequality, we get that
2 1 ( x + y ) 2 + 2 1 ( x − y ) 2 + 1 ≥ 2 1 ( x + y ) 2 + 1 ≥ 2 ( x + y ) .
It remains to verify that equality can hold throughout, which gives us C 1 = 2 .
For the second inequality, we use the same trick. The inequality is equivalent to
4 3 ( x + y ) 2 + 4 1 ( x − y ) 2 + 1 ≥ C 2 ( x + y )
By AM-GM and the trivial inequality, we get that
4 3 ( x + y ) 2 + 4 1 ( x − y ) 2 + 1 ≥ 4 3 ( x + y ) 2 + 1 ≥ 3 ( x + y )
It remains to verify that equality can hold throughout, which gives us C 2 = 3 .
This approach allows you to easily solve for the general case where
x 2 + y 2 + A x y + 1 ≥ C ( x + y )
as long as 0 ≤ A ≤ 2 .
Log in to reply
Agh, outsmarted again.
At least my solution seems more general. I'm pretty sure it will work for any type of inequality of the sort.
Log in to reply
The methods are actually the same, and the underlying relation between the proofs is how AM-GM and the quadratic formula are related through the discriminant. The ∣ A ∣ ≤ 2 condition is crucial, in getting an actual bound after pushing through the work.
Generalizing your approach to x 2 + y 2 + A x y + 1 ≥ C ( x + y ) , we have the quadratic equation
x 2 + ( A y − C ) x + ( y 2 − C y + 1 ) ≥ 0
Since it cannot have distinct real roots, the discriminant is non-positive, so for all values of y ,
⇒ ( A y − C ) 2 − 4 ( y 2 − C y + 1 ) ≤ 0 ( A 2 − 4 ) y 2 + ( 4 C − 2 A C ) y + ( C 2 − 4 ) ≤ 0
For this to always be negative, the leading coefficient is non-positive. Hence, we must have A 2 − 4 ≤ 0 as claimed.
For such a value of A , taking the discriminant to be non-positive, we must have
⇒ ⇒ ( 4 C − 2 A C ) 2 − 4 ( A 2 − 4 ) ( C 2 − 4 ) ≤ 0 A 2 − A C 2 + 2 C 2 − 4 ≤ 0 ∣ C ∣ ≤ 2 + A
Thus, the maximum value of C is 2 + A . This agrees with the A = 0 , A = 1 case that we done above.
Now, show that that my method yields the same result.
@Finn Hulse this works :)
Log in to reply
I know. :D
YES! Thank you so much Daniel.
For the first part, x 2 + y 2 + 1 ≥ 2 ( x + y ) 2 + 1 ≥ 2 2 ( x + y ) 2 = 2 ( x + y ) . with equality when x = y and when ( x + y ) 2 = 2 both of which simultaneouly indicate towards ( x , y ) = 2 1 , 2 1
Obviously, we need to consider only both positive reals x,y to get a good lower bound becasue the LHS is always positive
I am pretty sure this was an extremely lucky guess. Now, I will post a solution, feel free to completely tear me apart for it.
Now, take x 2 + y 2 + 1 . If x y = 2 1 , then the expression can be written as ( x + y ) 2
Now, for the maximum, equality must hold. Now, ( x + y ) 2 = C 1 ( x + y ) ⇒ C 1 = x + y . By A.M. G.M. Inequality, the minimum value for x + y = 2 2 = 2 . Thus, C 1 = 2 .
Now, x 2 + y 2 + x y + 1 = ( ( x + y ) 2 + x y ) because x y is assumed to be 2 1 . Thus, the expression boils down to C 2 = 2 2 5 . Thus, C 1 C 2 = 2 5 = 4 2 5 ≈ 6
I am completely lost. This is a complete guess. Could someone else please post a solution?
The first part is right, but not the second inequality. It's actually 3 . But can you prove it? It's not much harder than the first one. :D
Log in to reply
The trouble with me trying to prove this is that it is not homogenous, and nobody likes non-homogenous inequalities :P
I'm working on it though. I think considering the two expressions as quadratics wrt x will work.
Correct the second part of ur proof. The second inequality has gone a little wrong
It should come as \sqrt{3}
Can we use Calculus?
lagrange multipliers is a tedious but guaranteed method to solve the problem...
For the first part, x 2 + y 2 + 1 ≥ 2 ( x + y ) 2 + 1 ≥ 2 2 ( x + y ) 2 = 2 ( x + y ) . with equality when x = y and when ( x + y ) 2 = 2 both of which simultaneouly indicate towards ( x , y ) = 2 1 , 2 1 . For the second part, x 2 + y 2 + x y + 1 = ( x + y ) 2 + 1 − x y ≥ ( x + y ) 2 + 1 − 4 ( x + y ) 2 = 4 3 ( x + y ) 2 + 1 ≥ 3 ( x + y ) . With equality only when x = y and ( x + y ) = 3 2 hold simultaneously which hints towards ( x , y ) = ( 3 1 , 3 1 ) .
First part:
expressing in terms of perfect squares:
(x - sqrt(c1/2))^2 + (y - sqrt(c1/2))^2 + 1 - [(c1^2)/2] > 0
for c1 to be maximum and have this condition to be true
1 - [(c1^2)/2]=0
c1 = sqrt(2)
we have c1 = sqrt(2)
For the second part: Since all terms are symmetrical wrt x and y .. Lets re-express [x^2 + y^2 + xy + 1 - c2(x+y)] this in terms of:[As sum of squares]
(ax + ay + b)^2 + (1 - a^2)*x^2 + (1-a^2)y^2 + (1-2a^2)xy + (1-b^2) - (c2+ 2ab)(x+y)
We have 3 degrees of freedom here Constraints: if we can vanish 'xy' in the second part of the expression we can regroup with the 2 degrees of freedom left to regroup the remaining to perfect square.
Vanishing xy we get a = 1/sqrt(2);
we get for the rest of the expression, leaving the perfect square;
1/2* [x^2 + y^2 +2* (1-b^2) - 2 (c2+ b sqrt(2))(x+y)]
to make this a sum of perfect squares we have the condition
(c2 + b*sqrt(2))^2 = (1-b^2)
evaluating c2 in terms of b and then applying condition for maximum of c2 we end up with b = sqrt(2/3); condition c2 = 2ab + b/2a = sqt(3)..
Not bad! Mind adding L A T E X ?
interpretazione analitica: mettendo l'uguale si ottengono l'equazione di una circonferenza e di un'ellisse (ruotata) con il >= ho i punti ad esse esterne. Ho applicato la condizione affinché esse si riducessero ad un solo punto
Via Google translate: analytic interpretation: putting the same you will get the equation of a circle and an ellipse (rotated) with the> = I have the points to them outside. I applied the condition so that they were reduced to a single point
english plz
Ottimo! Sono davvero felice di vedere un approccio geometrico a questo problema! Ti dispiacerebbe spiegare un po 'di più, preferibilmente in inglese?
Problem Loading...
Note Loading...
Set Loading...
Let's work on the first equation first.
x 2 + y 2 + 1 ≥ C 1 ( x + y )
Considering this as a quadratic wrt x we rearrange to get x 2 − C 1 x + ( y 2 − C 1 y + 1 ) ≥ 0
Using the quadratic formula, we get x = 2 C 1 ± − 4 y 2 + 4 C 1 y + ( C 1 2 − 4 )
Note that if the discriminant Δ > 0 , then there exists some x that does not satisfy the inequality. Thus, we must have Δ ≤ 0 , or − 4 y 2 + 4 C 1 y + ( C 1 2 − 4 ) ≤ 0
Now we have to solve a quadratic wrt y. Using the quadratic formula again, we see that y = 2 C 1 ± 2 C 1 2 − 4
Similarly to above, note that if Δ > 0 in this new quadratic, then there exists a y such that it does not satisfy the inequality; thus, Δ ≤ 0 , or 2 C 1 2 − 4 ≤ 0
Solving for C 1 , we get that C 1 ≤ 2 , so the maximum value of C 1 is 2 .
This also means that x = y = 2 C 1 is the equality case (since the discriminant in both cases disappears) so our equality case is ( x , y ) = ( 2 2 , 2 2 )
We have the equation x 2 + y 2 + x y + 1 ≥ C 2 ( x + y )
Using a similar idea to before, we express the inequality as a quadratic wrt x: x 2 + ( y − C 2 ) x + ( y 2 − C 2 y + 1 ) ≥ 0
Using the quadratic equation, we obtain x = 2 C 2 − y ± − 3 y 2 + 2 C 2 y + C 2 2 − 4
Similar to last time, we must have Δ ≤ 0 , so − 3 y 2 + 2 C 2 y + C 2 2 − 4 ≤ 0
Solving the new quadratic wrt y we have y = 6 2 C 2 ± 1 6 C 2 2 − 4 8
Like before, we must have Δ ≤ 0 , so 1 2 C 2 2 − 4 8 ≤ 0
Solving for C 2 gives C 2 ≤ 3 , so the maximum of C 2 is 3 .
This also means that the equality case is y = 3 C 2 , and x = 2 C 2 − y , which simplifies to ( x , y ) = ( 3 3 , 3 3 )
Finishing of the problem, we have C 1 C 2 = 2 ⋅ 3 = 6 , so our answer is 6 .