Given 81 variables that satisfy
0 ≤ a 1 ≤ a 2 ≤ … ≤ a 8 1 ≤ 1 ,
what is the maximum value of
[ 9 i = 1 ∑ 8 1 a i 2 ] + ⎣ ⎡ 1 ≤ j < k ≤ 8 1 ∑ ( a k − a j + 1 ) 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.
The decimal is misleading. I wasn't quite sure of my answer of 8532 but upon seeing the decimal warning, I decided to just click "reveal solutions". I feel like this is unfair. Thoughts?
After two wrong guesses, I came upon the same reasoning. I started with calculus, which led to a linear system of equations, which wasn't necessary but I did it anyway. But I realized because there is no stationary point that is a local maximum, the maximum has to occur on the boundary (because of continuity in the derivatives). So, after playing around with some combinations of 1's and 0's, and counting the number of 1's and 2's being squared in the [ s t u f f ] piece, I concluded, just as you did, that for some n , we must have a 1 = a 2 = . . . = a n = 0 and a n + 1 = a n + 2 = . . . = a 8 1 = 1 . This allowed me to simplify the expression to ( 8 1 − x ) ( 9 + 3 x ) + ( 8 1 ) ( 4 0 ) . However, like an idiot I assumed the maximum occurred when 8 1 − x = 9 + 3 x , which led to an answer of 7209. I don't understand how I could be so stupid!
Awesome problem by the way
Wonderful....question.
First assumption: Must be a bunch of 0s then a bunch or 1s.
Then let x be the number of 1s in the sequence. The sum is thus:
9 x - Add all the 1 2
4 x ( 8 1 − x ) - Add all the ( 1 − 0 + 1 ) 2
. 5 x ( x − 1 ) - Add all the ( 1 − 1 + 1 ) 2
. 5 ( 8 1 − x ) ( 8 1 − x − 1 ) - Add all the ( 0 − 0 + 1 ) 2
So the final equation is:
9 x + 4 x ( 8 1 − x ) + . 5 x ( x − 1 ) + . 5 ( 8 1 − x ) ( 8 1 − x − 1 )
Simple quadratic that can be simplified and the maximum found at the vertex.
Could you please tell me how to write LaTeX here like you did. I have the LaTeX writer but it gives off the pdf file, but how to write it here or anywhere?
Log in to reply
Just check the formatting guide when you are posting. and to enclose your math as instructed.
After squaring we get (1+2+3+...+80)+ 89((a1)²+(a2)²+...+(a81)²)+ 4×(40(a81-a1)+39(a80-a2)+ ...+2(a43-a39)+(a42-a40)) -2(a81(a1+a2+...+a80)+ a(80(a1+a2+...+a79)+ +...+a3(a1+a2)+a2×a1
So first we can conclude a81=a80=…=a42=1 a40=…=a2=a1=0 For a41 we try 0 or 1 and we see that sum is greater when a41=1
Now,we can calculate value as 80×81/2+89×41+4×(40+39+...+1) -2(40+39+...+1)=40×81+40×89+89+ +40×41=40×211+89=8529 But if we put x40=1 instead of 0 We'll get 40×81+42×89+4×(40+...+1) -4-2(41+...+1)= 40×81+40×89+178+41×40-4-82= 40×211+92=8440+92=8532 If we put a39=1 instead of zero we'll get 8352+89-2×42-4×2=8529 So max is 8352 when a1=a2=…=a39=0 a40=a41=…=a81=1
How did you get to "We can conclude that the top values should be 1 and the bottom values should be 0"? In particular, we are introducing 8 9 a 1 2 , so why isn't it good to set a 1 = 1 ?
Log in to reply
to maximize a81-a1,a80-a2...
Log in to reply
Well, to find the maximum of f ( x ) + g ( x ) + h ( x ) , we cannot simply try and maximize g ( x ) without considering what happens to f ( x ) and h ( x ) . For example, which I can be pretty convinced that we want a 8 1 − a 1 to be maximized, it is not clear to me that we also want a 4 0 = 0 , since that appears in f ( x ) as 8 9 a 4 0 2 and also in h ( x ) multiple times.
E.g. If we wanted to maximize ( 1 − x 2 ) + ( 1 − ( x − 1 ) 2 ) , we cannot simply say that "the maximum for the first term happens at x = 0 , so we use that value.
Log in to reply
@Calvin Lin – 89a1*a1-160a1-2a1(a2+a3+...+a80) is greater when a1=0 than a1=1
Log in to reply
@Nikola Djuric – you're right but i do it by intuition,i was aware it need to be proved,but i didn't findout how...
Log in to reply
@Nikola Djuric – Yup, the tricky part of the solution is to prove that you indeed have a maximum, as opposed to "state the solution yet that you think yields the maximum".
This is the first step towards understanding mathematical proofs for inequalities, so I applaud your bravery for exploring the unknown. Keep it up!
Numbers. Tons of numbers ~o~
Problem Loading...
Note Loading...
Set Loading...
(This is a sketch, with enough information for you to fill in the details.)
First, fix an integer k . Assume that all variables a i are fixed for i = k . We want to find the value of a k , subject to a k − 1 ≤ a k ≤ a k + 1 which would maximize the expression. It is clear that the expression is a quadratic in a k , with positive leading coefficient. Hence, the maximum occurs at the endpoints, which means that a k = a k − 1 or a k = a k + 1 .
Since this is true for all k , this implies that we must have 0 = a 1 = a 2 = … = a n , a n + 1 = a n + 2 = … = a 8 1 = 1 for some integer n from 0 to 81. The expression thus simplifies to
9 ( 8 1 − n ) + [ stuff ] ,
which is a quadratic in n (with negative leading coefficient). We can easily see that the maximum happens when n = 3 9 (yes, values were specially chosen to make the maximum occur at an integer), and this would give us the value 8532.