Difference between revisions of "2012 USAMO Problems/Problem 6"
(→Solution 2) |
(→Note:) |
||
Line 44: | Line 44: | ||
Since <math>2^n - 2^{\frac{n}{2}} \le n2^{n-3}</math> for all <math>n \ge 8</math> and <math>\sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \ge 0</math> for all <math>k</math> and <math>i</math>, it is apparent that: | Since <math>2^n - 2^{\frac{n}{2}} \le n2^{n-3}</math> for all <math>n \ge 8</math> and <math>\sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \ge 0</math> for all <math>k</math> and <math>i</math>, it is apparent that: | ||
<cmath>\sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=0}^{k-1} \binom{\frac{n}{2}}{i} \le n2^{n-3}</cmath> | <cmath>\sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=0}^{k-1} \binom{\frac{n}{2}}{i} \le n2^{n-3}</cmath> | ||
− | must be true for all <math>n \ge 8</math>(because if we rewrite this we get <math>2^{\frac{n}{2}}\left(8-n\right) \le 8</math>. For all <math>2 \le n < 8</math> however, we may use some logic to first layout a plan. Since for <math>n=6,n=4,</math> and <math>n=2</math>, <math>2^{n} - 2^{\frac{n}{2}} = n2^{n-3} + 1</math>, we may say that whole sum will be less than <math>n2^{n-3}</math> because <math>\sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \ge 1</math> for all <math>n \ge 2</math> Plugging this inequality back in gives us: | + | must be true for all <math>n \ge 8</math>(because if we rewrite this we get <math>2^{\frac{n}{2}}\left(8-n\right) \le 8</math>.) For all <math>2 \le n < 8</math> however, we may use some logic to first layout a plan. Since for <math>n=6,n=4,</math> and <math>n=2</math>, <math>2^{n} - 2^{\frac{n}{2}} = n2^{n-3} + 1</math>, we may say that whole sum will be less than <math>n2^{n-3}</math> because <math>\sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \ge 1</math> for all <math>n \ge 2</math> Plugging this inequality back in gives us: |
<cmath>2^{n} - 2^{\frac{n}{2}} - \sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \le 2^{n} - 2^{\frac{n}{2}} - 1 = n2^{n-3}</cmath> | <cmath>2^{n} - 2^{\frac{n}{2}} - \sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \le 2^{n} - 2^{\frac{n}{2}} - 1 = n2^{n-3}</cmath> | ||
because of the fact that <math>- \sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \le -1</math> | because of the fact that <math>- \sum_{k=1}^{\frac{n}{2}} \binom{\frac{n}{2}}{k} \sum_{i=k}^{\frac{n}{2}} \binom{\frac{n}{2}}{i} \le -1</math> |
Revision as of 21:16, 13 December 2022
Contents
Problem
For integer , let , , , be real numbers satisfying For each subset , define (If is the empty set, then .)
Prove that for any positive number , the number of sets satisfying is at most . For what choices of , , , , does equality hold?
Solution 1
For convenience, let .
Note that , so the sum of the taken two at a time is . Now consider the following sum:
Since , it follows that at most sets have .
Now note that . It follows that at most half of the such that are positive. This shows that at most sets satisfy .
Note that if equality holds, every subset of has . It immediately follows that is a permutation of . Since we know that , we have that .
Solution 2
Let It is evident that for evens because of the second equation and for odds(one term will be 0 to maintain the first condition). We may then try and get an expression for the maximum number of sets that satisfy this which occur when : since it will be for any choice of A we pick, it will have to be greater than which means we can either pick 0 negative or negatives for j positive terms which gives us: and For odd values, let it be the same as the last even valued sequence where n is even(i.e. the same as the sequence before it but with an extra 0 in one of the spots). Then, the following is apparent: Thus, we may say that this holds to be true for all since grows faster than the sum. Note that equality holds when for all i which occurs when and since is the only choice for
Note:
The proof to the last inequality is as follows: First we may rewrite this as being: Thus, (the second equation is because the sum of the binomial coefficient is ) Since for all and for all and , it is apparent that: must be true for all (because if we rewrite this we get .) For all however, we may use some logic to first layout a plan. Since for and , , we may say that whole sum will be less than because for all Plugging this inequality back in gives us: because of the fact that
See Also
2012 USAMO (Problems • Resources) | ||
Preceded by Problem 5 |
Last Problem | |
1 • 2 • 3 • 4 • 5 • 6 | ||
All USAMO Problems and Solutions |
The problems on this page are copyrighted by the Mathematical Association of America's American Mathematics Competitions.