8

I stumbled upon the following identity $$\sum_{k=0}^n(-1)^kC_k\binom{k+2}{n-k}=0\qquad n\ge2$$ where $C_n$ is the $n$th Catalan number. Any suggestions on how to prove it are welcome!

This came up as a special case of a generating function for labeled binary trees. Actually I can directly prove the identity is zero by showing that certain trees don't exist, but I expect that seeing a direct proof will help me find nice closed formulae for other coefficients of the generating function.

2 Answers2

5

My answer uses the same methods as that of tc2718's answer to a very similar problem.

Let $a_n$ be the quantity in question, and let $F(x)=\sum_{n\ge 0}a_nx^n$. We will evaluate $F$ by switching the order of summation, applying the binomial theorem to the inner sum, rearranging a bit, and then recognizing the expression as the Catalan generating function evaluated at $-x-x^2$. $$ \begin{align} F(x) &=\sum_{n\ge 0}\sum_{k=0}^n (-1)^k C_k \binom{k+2}{n-k}x^n \\&=\sum_{k=0}^\infty (-1)^k C_k \sum_{n= k}^\infty\binom{k+2}{n-k}x^n \\&=\sum_{k=0}^\infty (-1)^k C_k \; x^k(1+x)^{k+2} \\&=(1+x)^2\sum_{k=0}^\infty C_k \; (-x-x^2)^k \\&=(1+x)^2 \frac{1-\sqrt{1-4(-x-x^2)}}{2(-x-x^2)} \\&=(1+x)^2 \frac{1-\sqrt{(1+2x)^2}}{2(-x-x^2)} \end{align} $$ Here, we must choose the branch of $\;\sqrt{\cdot}\;$ where $\sqrt{(1+2x)^2}=+(1+2x)$, since the other choice causes a pole at $x=0$, and our definition of $F(x)$ makes it clear there is no pole there. Finally, \begin{align} F(x)=(1+x)^2 \cdot \frac{1-(1+2x)}{-2x(1+x)}=1+x \end{align}

This closed form expression for $F(x)=\sum_{n\ge 0}a_nx^n$ shows that $a_0=a_1=1$, while $a_n=0$ for $n\ge 2$.

Mike Earnest
  • 75,930
5

This answer uses Umbral calculus.

Given the formal variable $\,c,\,$ define the formal power series linear operator

$$ L\!\!\left[\sum_{n=0}^\infty a_nc^n\right] := \sum_{n=0}^\infty a_nC_n. $$

Define the Catalan number generating function

$$ C(x) := L\left[\frac1{1-c\,x}\right] = \sum_{n=0}^\infty C_n x^n = \frac{1-\sqrt{1-4x}}{2x}. $$

Define the generating function

$$ F(x) := \sum_{n=0}^\infty (-1)^kC_k\binom{k+2}{n-k}x^n. $$

Define the generating function

$$ f(x,y) := \sum_{n=0}^\infty \left(\sum_{k=0}^n (-1)^k{k+2\choose n-k}y^k\right)x^n. $$

Now, by previous definitions

$$ F(x) = L[f(x,c)]. $$

Verify that

$$ f(x,y) = \frac{(1 + x)^2}{ 1 + y (x + x^2) }. $$

Now get

$$ F(x) = (1+x)^2 C(-(x+x^2)) = 1+x $$

from

$$ C(-(x\!+\!x^2)) \!=\! \frac{1\!-\!\sqrt{1\!+\!4(x\!+\!x^2)}} {-2(x\!+\!x^2)} \!=\! \frac{1-(1\!+\!2x)}{-2(x\!+\!x^2)} \!=\! \frac1{1\!+\!x}. $$


Notice, my answer to MSE question 4317353 also uses Umbral calculus to prove another Catalan number recursion identity. Also, the accepted answer to this question is very similar to mine except it does not use the Umbral calculus linear operator.

Somos
  • 35,251
  • 3
  • 30
  • 76
  • 1
    (+1) Extremely slick! As a quick question, I'm familiar with the umbral calculus, but not particularly comfortable with it. Do you have any insight as to which problems are amenable to the technique, and do you know of any resources for practicing? Ideally there would be something like Generatingfunctionology but for the umbral calculus, though I know that's a high bar to clear! – HallaSurvivor Dec 15 '21 at 19:03
  • 1
    @HallaSurvivor That is a good question! Please ask this as a question here on MSE and you may get good answers for it. – Somos Dec 15 '21 at 19:07
  • I'd be happy to ^_^ – HallaSurvivor Dec 15 '21 at 19:22
  • If you have some time to chime in, I've asked the question here – HallaSurvivor Dec 15 '21 at 19:51
  • Cool. This looks like a very powerful technique which I look forward to diving into further! – Cheerful Parsnip Dec 17 '21 at 01:58
  • @HallaSurvivor The book by Roman cited here is a nice introduction to Umbral Calculus. It includes many illuminating examples and is highly accessible. – Bill Dubuque Dec 17 '21 at 11:27