-1
$\begingroup$

Does anyone know how to show the following combinatorial equality, $\sum_{i=0}^{n}\left(n-i\right)^{2}\binom{2n}{i}=n\cdot4^{n-1}$?

By the way, this is not a homework problem, otherwise one would be able to search the answer.

Thanks.

$\endgroup$
1
  • 4
    $\begingroup$ As Robin Chapman mentions, your sum is half of the full sum $\sum_{i=0}^{2n}(n-i)^2\dbinom{2n}i$. Note that $\sum_{i=0}^{2n}\dbinom{2n}ix^i=(1+x)^{2n}$ and if you aplly the differential operator $x\dfrac{d}{dx}$, you also have the closed formulas for $\sum_{i=0}^{2n}i^l\dbinom{2n}ix^i$ with $l=1,2$. A linear combination of the three will allow you to compute your expression. It's a homework problem! $\endgroup$ May 23, 2010 at 9:50

3 Answers 3

2
$\begingroup$

It's half the sum of the same thing from $0$ to $2n$, which in turn is easily related to the variance of the number of heads in a sequence of $2n$ tosses of a fair coin.

$\endgroup$
1
  • 1
    $\begingroup$ Robin, I spent time following your combinatorial hint. If the author claims that his problem isn't a homework, your combinatorial solution is definitely best possible. +1 $\endgroup$ May 23, 2010 at 11:59
0
$\begingroup$

Try the Wilf-Zeilberger method and its friends. This automatically proves many such (hypergeometric) identities. See the book A = B

http://www.math.upenn.edu/~wilf/AeqB.html

$\endgroup$
0
$\begingroup$

I suggest you take a look on hypergeometric series.

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.