Real numbers a 1 , a 2 , … , a 8 0 0 9 0 satisfy
( a 1 + a 2 + … + a 8 0 0 9 0 ) 2 + a 1 2 + a 2 2 + … + a 8 0 0 9 0 2 ≤ 8 0 0 9 0 .
What is the smallest integer value of N such that for all possible such tuples ( a 1 , . . . , a 8 0 0 9 0 ) , we have ∣ a i ∣ ≤ N for all i ?
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.
Nice solution. An alternative would be to use the RMS-AM inequality:
a 2 2 + a 3 2 + … + a n 2 ≥ n − 1 ( a 2 + a 3 + … + a n ) 2
Let s = a 2 + … + a n . This gives:
n − 1 ≥ ( a 1 + … + a n ) 2 + a 1 2 + … + a n 2 ≥ ( a 1 + s ) 2 + a 1 2 + n − 1 s 2 .
Now express this as a quadratic equation in s - the fact that s is real gives a discriminant condition in a 1 .
My solution goes like this:
A certain inequality (I forgot the name) holds that:
a 1 2 + a 2 2 + ... + a 8 0 0 9 0 2 ≥ 8 0 0 9 0 ( a 1 + a 2 + . . . + a 8 0 0 9 0 ) 2
By Addition Property of Inequality:
a 1 2 + a 2 2 + ... + a 8 0 0 9 0 2 + ( a 1 + a 2 + . . . + a 8 0 0 9 0 ) 2 ≥ 8 0 0 9 0 8 0 0 9 1 ( a 1 + a 2 + . . . + a 8 0 0 9 0 ) 2
Using the given above:
80090 ≥ a 1 2 + a 2 2 + ... + a 8 0 0 9 0 2 + ( a 1 + a 2 + . . . + a 8 0 0 9 0 ) 2 ≥ 8 0 0 9 0 8 0 0 9 1 ( a 1 + a 2 + . . . + a 8 0 0 9 0 ) 2
By Squeezing Method:
80090 ≥ 8 0 0 9 0 8 0 0 9 1 ( a 1 + a 2 + . . . + a 8 0 0 9 0 ) 2
By Multiplication Property of Inequality ( 8 0 0 9 0 > 0 and 8 0 0 9 1 > 0 so there is nothing to worry about the inequality symbol) :
8 0 0 9 1 8 0 0 9 0 2 ≥ ( a 1 + a 2 + . . . + a 8 0 0 9 0 ) 2
Taking the square root of both sides:
284 ≥ a 1 + a 2 + . . . + a 8 0 0 9 0
Therefore any a i is always less than 284
Log in to reply
This is my solution. The calculator only displayed 283 when i took the square root of both sides of the inequality that's why I arrived to a wrong answer. Thanks
a's are reals, not necessarily positive!? I use the easily proven fact that a1 is maximized when a2 through a80090 are identical negative numbers (each = -t ), so that a1^2+ 80089t^2+ (a1-80089t)^2<=80090. In this quadratic form, a1 is maxed when a1=80090t. Plug in and solve for a1 -> a1<284.
Oh I think it is the Cauchy–Schwarz inequality :)
Almost did this correctly, it was that nasty calculator rounded the value down to 283
This follows from two claims:
Step 1 : If
( a 1 + a 2 + … + a 8 0 0 9 0 ) 2 + a 1 2 + a 2 2 + … + a 8 0 0 9 0 2 ≤ 8 0 0 9 0 ( ∗ ) ,
then ∣ a i ∣ ≤ 2 8 3 . 9 for all i .
Proof : If ∣ a i ∣ > 2 8 3 . 9 , then a i 2 > 8 0 0 9 0 so the inequality (*) does not hold. (QED)
Step 2 : There exist real numbers a 1 , … , a 8 0 0 9 0 such that (*) holds and a 1 = 2 8 3 .
Proof : Indeed, let a 1 = 2 8 3 and a 2 = a 3 = … = a 8 0 0 9 0 = − 2 8 3 1 . Then the expression in (*) is exactly 80090. (QED)
I too went by this method but I wasn't sure about it.
My apologies, I just realised that my solution only proves that N < 2 8 2 is not viable.
It takes more effort to prove that ∣ a i ∣ ≤ 2 8 3 for all i . Will think more about it. :(
I'm starting to think the solution is wrong. Correct me if I'm wrong, but:
seem to satisfy the inequality, with ∣ a 1 ∣ > 2 8 3 ? That would mean N = 2 8 3 fails.
We need to find the largest a i ever possible to find the smallest N. We can accomplish this by making all but one a i 0. This simplifies the inequality to a 2 ≤ 8 0 0 9 0 .
From this we know that the highest possible a is 8 0 0 9 0 ≈ 2 8 3 . 0 0 1 8 . The integer immediately above is 284.
N must be 284 .
If all but one are zero, the inequality becomes 2 a 2 ≤ 8 0 0 9 0 , giving ∣ a ∣ ≤ 2 0 0 . 1 1 .
Log in to reply
I see... you're right. How did my answer work, then? Was that a coincidence?
Log in to reply
Yes. If you look at my answer, you see that the upper bound of 8 0 0 9 1 8 0 0 9 0 is less than 8 0 0 9 0 , but not (quite!) enough to take the integer answer down to 2 8 3 .
Since the inequality works with very very small a i, the largest a i, a k satisfies 284>a k>283
Log in to reply
@Cuong Doan – There is tension in the inequality though. If you use very small negative a i , then the sum can still be very large. You need to justify that statement.
I see... you're right. How did my answer work, then? Was that a coincidence?
@Darryl Y,
You're off to a good start, but you need to consider this question: if one of the a's is about 283, what do you want the other a's to be (given that you want the expression in parentheses to be about 0)?
Problem Loading...
Note Loading...
Set Loading...
We wish to maximise f ( x ) = x 1 subject to the relation g ( x ) = ( x 1 + x 2 + ⋯ + x n ) 2 + x 1 2 + x 2 2 + ⋯ + x n 2 = n By Lagrange Multipliers, we need to find λ such that ∇ f − λ ∇ g = 0 . For 2 ≤ i ≤ n , the i th component of ∇ f − λ ∇ g is − λ [ 2 ( x 1 + x 2 + ⋯ + x n ) + 2 x i ] = 0 so we deduce that x 2 = x 3 = ⋯ = x n = − y where x 1 = n y . Considering the first component of ∇ f − λ ∇ g would determine λ , but we are not interested in that. We have n = g ( x ) = = ( x 1 + x 2 + ⋯ + x n ) 2 + x 1 2 + x 2 2 + ⋯ + x n 2 y 2 + ( n y ) 2 + ( n − 1 ) y 2 = n ( n + 1 ) y 2 and hence y = n + 1 1 . Thus x 1 = n + 1 n .
Thus it follows that ∣ x i ∣ ≤ n + 1 n for all 1 ≤ i ≤ n whenever g ( x ) = n .
Now 8 0 0 8 9 = 2 8 3 2 and it is easy to show that m < m 2 + 2 m 2 + 1 < m + 1 and hence the smallest integer value of N that we want to answer the question is ⌊ 8 0 0 9 1 8 0 0 9 0 ⌋ + 1 = 2 8 4 The answer is not 2 8 3 , but calculators will not spot this, since 8 0 0 9 1 8 0 0 9 0 = 2 8 3 . 0 0 0 0 0 0 0 2 2 .