Proving the Combinatorial Identity for B_k(x)B_l(x)

  • Thread starter Thread starter BrownianMan
  • Start date Start date
  • Tags Tags
    Identity
BrownianMan
Messages
133
Reaction score
0
Let B_{k}(x)=\sum_{n\geq 0}\binom{n}{k}\frac{x^{n}}{n!}. Show that

B_{k}(x)B_{l}(x)=\frac{1}{2^{k+l}}\binom{k+l}{l}B_{k+l}(2x).

I'm having some trouble with this one. Does anyone have any hints? I've tried using Cauchy product and Chu-Vandermonde equality but I get B_{k}(x)B_l(x)=\sum_{n\geq 0}\binom{n}{k+l}\frac{(2x)^{n}}{n!}=B_{k+l}(2x).

I'm not sure where the \frac{1}{2^{k+l}}\binom{k+l}{l} comes from.

I'm also suppose to take the x^n/n! coefficient of both sides and show which binomial coefficient identity this gives.
 
Physics news on Phys.org
BrownianMan said:
Let B_{k}(x)=\sum_{n\geq 0}\binom{n}{k}\frac{x^{n}}{n!}. Show that

B_{k}(x)B_{l}(x)=\frac{1}{2^{k+l}}\binom{k+l}{l}B_{k+l}(2x).

I'm having some trouble with this one. Does anyone have any hints? I've tried using Cauchy product and Chu-Vandermonde equality but I get B_{k}(x)B_l(x)=\sum_{n\geq 0}\binom{n}{k+l}\frac{(2x)^{n}}{n!}=B_{k+l}(2x).

I'm not sure where the \frac{1}{2^{k+l}}\binom{k+l}{l} comes from.

I'm also suppose to take the x^n/n! coefficient of both sides and show which binomial coefficient identity this gives.
Probably most useful to you is to find the error in what you have done, but for that you'll need to post your working. (And a statement or URL for any existing result your rely on.)
But fwiw, I get:
B_{k}(x)=\sum_{n\geq 0}\binom{n}{k}\frac{x^{n}}{n!} = \sum_{n\geq k}\binom{n}{k}\frac{x^{n}}{n!}<br /> = \sum_{n\geq 0}\binom{n+k}{k}\frac{x^{n+k}}{(n+k)!}<br /> = x^k\sum_{n\geq 0}\frac{x^{n}}{n!k!}<br /> = \frac{x^k}{k!}\sum_{n\geq 0}\frac{x^{n}}{n!} = \frac{x^k}{k!}e^x
 
I don't see how Bk(x) = x^k/k! * e^x. I checked with specific values and it doesn't hold.
 
BrownianMan said:
I don't see how Bk(x) = x^k/k! * e^x. I checked with specific values and it doesn't hold.
What values? Seems ok for k = 0, 1, 2.
 
haruspex said:
What values? Seems ok for k = 0, 1, 2.

You're right. My mistake.

Thanks.
 
OK, good. Did that help you solve the OP?
 
Yes, I solved it. Thanks again!
 
Back
Top