0
$\begingroup$

EDIT: Just to be clear ${n+k \brace n}$ denotes the (n+k,n)-th stirling number of the second kind.

I have been trying to find a decent form for the generating function $\sum_{n=0}^\infty{n+k \brace n}x^n$ but it has remained largely untenable. There doesn't appear to be any useful recurrence here and any attempts at a combinatorial argument are made difficult by the fact that Stirling numbers primarily play well with exponential generating functions. Even when I do try to look at the exponential generating function it doesn't help because neither of the inputs to ${- \brace -}$ are fixed. I was hoping that somebody could help me out with making this generating functions more workable.

$\endgroup$
2
  • $\begingroup$ Calculate the first few values of $k$ and have a look at this grid of numbers. oeis.org/A008517 $\endgroup$ Commented Dec 8 at 22:15
  • $\begingroup$ Please consider defining the expression ${n+k \brace n}$. $\endgroup$ Commented yesterday

1 Answer 1

3
$\begingroup$

the $n$th diagonal of either Stirling triangle (since ${n\brace k}={-k\brack-n}$) is a polynomial of degree $2n$ (comes out of the diagonal hockey stick identities obtainable from the recurrence; see this answer), which means its o.g.f. is rational, with denominator of degree $2n+1$.

as mentioned by @DonaldSplutterwit, we have the dual identities $$\begin{aligned} {n+k\brace k}&=\sum_{i=0}^n\left(\!{2n+1\choose k-i}\!\right)\left\langle\!\left\langle{n\atop i-1}\right\rangle\!\right\rangle\\ {n+k\brack k}&=\sum_{i=0}^n\left(\!{2n+1\choose k-i}\!\right)\left\langle\!\left\langle{n\atop n-i}\right\rangle\!\right\rangle \end{aligned}$$which come out of$$\begin{aligned} \sum_{k=0}^\infty{n+k\brace k}x^k&=\frac{\sum_{i=0}^n\left\langle\!\left\langle{n\atop i-1}\right\rangle\!\right\rangle x^i}{(1-x)^{2n+1}}\\ \sum_{k=0}^\infty{n+k\brack k}x^k&=\frac{\sum_{i=0}^n\left\langle\!\left\langle{n\atop n-i}\right\rangle\!\right\rangle x^i}{(1-x)^{2n+1}} \end{aligned}$$

where $\left(\!{n\choose k}\!\right)=(-1)^k{-n\choose k}={n+k-1\choose k}$, and the double-angle-bracketed coefficients the second-order Eulerian numbers according to the indexing convention of $\left\langle\!\left\langle{n\atop k}\right\rangle\!\right\rangle=$A201637(n,k) specified by Concrete Mathematics on p. 270.

these can be implemented easily via $\left\langle\!\left\langle{n\atop k}\right\rangle\!\right\rangle=(k+1)\left\langle\!\left\langle{n-1\atop k}\right\rangle\!\right\rangle+(2n-1-k)\left\langle\!\left\langle{n-1\atop k-1}\right\rangle\!\right\rangle$, which can be derived straightforwardly by by writing the defining o.g.f. relation the other way around and using annihilated coefficient extractors, Zeilberger's algorithm (in particular this implementation handling sums involving Stirling numbers) or (quite laborious) arithmetic manipulation (performed here).

$\endgroup$

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.