Define the double factorial as being a continous, non-hybrid function

Join the discussion
Ask a follow-up here, or get your own question answered by working scientists, mathematicians and engineers — people, not an autocomplete.
Real named experts · corrections over time · the nuance an AI answer skips
4 replies · 2K views
Saracen Rue
Messages
150
Reaction score
10
TL;DR
Define ##x!!## as being a continuous, non-hybrid function over the domain of ##[-1, \infty)##
The double factorial, ##n!## (not to be confused with ##(n!)!##), can be defined for positive integer values like so:
$$n!=n(n−2)(n−4)(n−6)...(n-a)$$
Where ##(n−a)=1## if ##n## is odd or ##(n−a)=2## if ##n## is even. Additionally, the definition of the double factorial extends such that ##(-1)!=0!=1##

I'm curious as to if it would be possible to definite a function which not only accurately evaluates all integer values for ##n!##, ##n \geq -1##, but also allows you to calculate the value of any real value of ##n## over the domain ##[-1, \infty)##
 
on Phys.org
Will be a discontinuity in second and higher derivatives at n=1. Therefore, single (not piece-wise) interpolation is not possible.
 
You can also continuously extend a function defined on a discrete set, but here's a concrete approach. I assume you're familiar with the gamma function ##\Gamma(z)=\int_0^{\infty} x^{z-1}e^{-x}dx## which satisfies ##\Gamma(n)=(n-1)!## when ##n## is a nonnegative integer.

There is the duplication formula ##\Gamma(z)\Gamma(z+1/2)=2^{1-2z}\sqrt{\pi} \ \Gamma(2z)##. When ##z=n## is a non-negative integer, this gives

##\Gamma(n+1/2)=\sqrt{\pi} \ 2^{1-2n}\frac{(2n-1)!}{(n-1)!}=\sqrt{\pi} \ 2^{-n} \frac{(2n-1)!}{2^{n-1}(n-1)!}=\sqrt{\pi} \ 2^{-n} (2n-1)!##

Rearranging, we find that the function ##\frac{1}{\sqrt{\pi}}2^{(1+z)/2}\Gamma(z/2+1)## is equal to ##z!## when ##z## is an odd (positive) integer. As far as I can tell, this is the standard analytic continuation of ##(2n-1)!##. Unfortunately, this formula is false when ##z## is even.

I found a stackexchange answer (https://math.stackexchange.com/ques...ion-of-factorial-be-done-for-double-factorial)
which gives the function ##2^{(1+2z-\cos(\pi z))/4}\pi^{(\cos(\pi z)-1)/4}\Gamma(z/2+1).## When ##z## is odd, this is just the function I gave above. When ##z=2n## is even, it gives ##2^nn!=(2n)!##. I don't think this is standard though.
 
Last edited:
  • Like
Likes   Reactions: Saracen Rue
Infrared said:
Unfortunately, this formula is false when ##z## is even.
For even ##z## one could use ##(2n)!=2^nn!##. A similar easy function is ##(2n-1)!=\dfrac{(2n)!}{2^nn!} ##. That leaves us with the problem to find one function for both cases, in which case we could simply substitute the gamma function.
 
  • Like
Likes   Reactions: Saracen Rue and dextercioby
Infrared said:
You can also continuously extend a function defined on a discrete set, but here's a concrete approach. I assume you're familiar with the gamma function ##\Gamma(z)=\int_0^{\infty} x^{z-1}e^{-x}dx## which satisfies ##\Gamma(n)=(n-1)!## when ##n## is a nonnegative integer.

There is the duplication formula ##\Gamma(z)\Gamma(z+1/2)=2^{1-2z}\sqrt{\pi} \ \Gamma(2z)##. When ##z=n## is a non-negative integer, this gives

##\Gamma(n+1/2)=\sqrt{\pi} \ 2^{1-2n}\frac{(2n-1)!}{(n-1)!}=\sqrt{\pi} \ 2^{-n} \frac{(2n-1)!}{2^{n-1}(n-1)!}=\sqrt{\pi} \ 2^{-n} (2n-1)!##

Rearranging, we find that the function ##\frac{1}{\sqrt{\pi}}2^{(1+z)/2}\Gamma(z/2+1)## is equal to ##z!## when ##z## is an odd (positive) integer. As far as I can tell, this is the standard analytic continuation of ##(2n-1)!##. Unfortunately, this formula is false when ##z## is even.

I found a stackexchange answer (https://math.stackexchange.com/ques...ion-of-factorial-be-done-for-double-factorial)
which gives the function ##2^{(1+2z-\cos(\pi z))/4}\pi^{(\cos(\pi z)-1)/4}\Gamma(z/2+1).## When ##z## is odd, this is just the function I gave above. When ##z=2n## is even, it gives ##2^nn!=(2n)!##. I don't think this is standard though.
Thank you, a function like ##f(z)=2^{(1+2z-\cos(\pi z))/4}\pi^{(\cos(\pi z)-1)/4}\Gamma(z/2+1)## is exactly what I was looking for. I wonder if it would be able to define a similar such function for the tripple factorial.