3

I'm reading the Wikipedia article on analytic continuation and I got stuck in the last step of the worked example. How to get the following result? $$\sum_{n=0}^{\infty}(-1)^n{n\choose k}(a-1)^{n-k}=(-1)^ka^{-k-1}$$

Obviously the sum starts from $n=k$ because ${n\choose k}=0$ for all $k>n$. I have tried by calculating some first terms of the sum and using the identity $${n\choose k}={n-1\choose k-1}+{n-1\choose k}$$ to see if there would be a useful pattern but I have had no success. Could someone tell how to arrive to the rhs from the lhs? This question is about the same exmple and there is no answer to the op's comment. So here is the comment as a proper question.

EDIT: I had a typo on the lhs: it's actually $(-1)^n$ instead of $(-1)^k$.

Thanks to Mike Earnest's answer suggesting Newton's Binomial Theorem I calculated $$(x+1)^{-n}=\sum_{i=0}^{\infty}{-n\choose i}x^i.$$ By simplifying the binomial coefficients $$\frac{-n(-n-1)(-n-2)...(-n-i+1)}{i!}=(-1)^i\frac{n(n+1)(n+2)...(n+i-1)}{i!}=(-1)^i\frac{n+i-1}{i!(n-1)}=(-1)^i{n-1 +i\choose i}=(-1)^i{n-1+i\choose n-1}$$ and then choosing $x=(a-1)$ gives: $$\sum_{n=k}^{\infty}(-1)^n{n\choose k}(a-1)^{n-k}=a^{-k-1}.$$ So I'm only missing the $(-1)^k$. Where can that be found?

clathratus
  • 17,161
jvkloc
  • 261

2 Answers2

2

We obtain \begin{align*} \color{blue}{\sum_{n=0}^{\infty}}&\color{blue}{(-1)^n{n\choose k}(a-1)^{n-k}}\\ &=\sum_{n=k}^{\infty}(-1)^n\binom{n}{k}(a-1)^{n-k}\tag{1}\\ &=\sum_{n=0}^{\infty}(-1)^{n+k}\binom{n+k}{k}(a-1)^{n}\tag{2}\\ &=(-1)^k\sum_{n=0}^{\infty}\binom{-k-1}{n}(a-1)^{n}\tag{3}\\ &=(-1)^k(1+(a-1))^{-k-1}\tag{4}\\ &\,\,\color{blue}{=(-1)^ka^{-k-1}} \end{align*} and the claim follows.

Comment:

  • In (1) we start with index $n=k$ since $\binom{n}{k}=0$ if $n<k$.

  • In (2) we shift the index to start with $n=0$ again.

  • In (3) we use the binomial identities $\binom{p}{q}=\binom{p}{p-q}$ and $\binom{-p}{q}=\binom{p+q-1}{q}(-1)^q$. We get \begin{align*} \binom{n+k}{k}=\binom{n+k}{n}=\binom{-(n+k)+n-1}{n}(-1)^n=\binom{-k-1}{n}(-1)^n \end{align*}

  • In (4) we use the binomial series expansion.

Markus Scheuer
  • 108,315
1

You need to use this general fact: $$ \sum_{n= 0}^\infty \binom nkx^n = \frac{x^k}{(1-x)^{k+1}}\qquad (|x|<1) $$ In your case, you want $x=a-1$.

To prove this, let $S_k=\sum_{n\ge 0}\binom nkx^k$. Using Pascal's identity, $\binom nk=\binom{n-1}{k-1}+\binom {n-1}k$, you can split this into two sums, and derive the equation $$ S_k = xS_{k-1}+xS_k, $$ which, combined with the base case $S_0=(1-x)^{-1}$, quickly leads to a proof by induction that $S_k=x^k(1-x)^{-(k+1)}$.

This also follows from Newton's binomial theorem: $$ \sum_{n\ge 0}\binom n{n-k}x^n =\sum_{n\ge 0}(-1)^{n-k}\binom{-(k+1)}{n-k}x^n =x^k\sum_{n\ge 0}\binom{-(k+1)}{n-k}(-x)^{n-k}=x^k(1-x)^{-(k+1)} $$

Mike Earnest
  • 75,930
  • It looks to me that the "general fact" is a strange one. $0=rhs$ for all $n<k$ and for $n=k$ it says $x^k=rhs$. Doesn't look correct to me. (If it is true, then it leaves an unwanted minus sign on the rhs. $1-x=-a$ for $x=a-1$.) – jvkloc Nov 16 '22 at 18:39
  • Wikipedia agrees it is correct, see second equation in "Ordinary generating functions" section: https://en.wikipedia.org/wiki/Binomial_coefficient#Ordinary_generating_functions Your complaint makes no sense; there is no "when $n=k$", it is a summation over all $n$. – Mike Earnest Nov 16 '22 at 19:08
  • True, it's a sum over all of them. I'll see the links, thank you. Only the minus sign remains a problem. Or is it also something my tired brain misreads/misunderstands? – jvkloc Nov 16 '22 at 19:10
  • Several proofs here: https://math.stackexchange.com/q/1941603/177399 – Mike Earnest Nov 16 '22 at 19:14
  • @Jonne You copied down the formula from the analytic continuation article wrong. It is supposed to be $(-1)^n$ in your summation, not $(-1)^k$. This is why you were unable to complete the proof. – Mike Earnest Nov 16 '22 at 19:19
  • 1
    I edited the question, it is correct now. Thanks again. I need to go to bed now, I will return to this tomorrow. – jvkloc Nov 16 '22 at 19:22