How do you take factorials of a fraction?
\[\left( \frac{ 1 }{ 2 } \right)!, \left( \frac{ 3 }{ 2 } \right)!\] etc @ganeshie8 Trying to understand this for gamma functions haha
\(\color{#000000 }{ \displaystyle \Gamma (x)=\int\limits_{0}^{\infty}z^{x-1}e^{-z}~dz }\) If I recall correctly, and then you can express the integral as a power series, and use partial sums to evaluate.
Do you know Euler reflection formula ?
\[\Gamma(1-z)\Gamma(z) = \dfrac{\pi}{\sin(\pi z)}\] for \(0\lt z\lt 1\)
plugin \(z=\frac{1}{2}\)
That's right, but I'm not sure how to take factorials of fractions, that's the part I'm not entirely sure about. \[\Gamma (n) = (n-1)!\]
Ooh interesting
correct me if i am wrong, but isn't \(\Gamma(n)=(n-1)!\) ?
right, euler reflection formula and the definition of factorial as recurrence relation allow you to compute the value of factorual of any fraction
I'm still not sure how that helps
and \[\Gamma(\frac{1}{2})\] is something goofy like \(\sqrt{\pi}\)
that is about all i remember
\[\Gamma (3/2) = (3/2-1)! = 1/2! = ?\]
I mean is there a way to do that without integrating
recurrence relation gives you \[\Gamma(x+1) = x\Gamma(x) \]
i guess we could try to compute the integral \[\int_0^{\infty}x^{-\frac{1}{2}}e^{-x}dx\]
without doing it my guess is parts, but i could be way off
\[\Gamma(\frac{3}{2}) = \Gamma(\frac{1}{2}+1) =\frac{1}{2}\Gamma(\frac{1}{2}) \]
You can find the value of \(\Gamma(\frac{1}{2})\) using euler reflection formula easily
\[\Gamma(1-z)\Gamma(z) = \dfrac{\pi}{\sin(\pi z)}\] plugin \(z=\frac{1}{2}\)
ok then what?
oh nvm
has a fit of stupidity there, yes, that works nicely for \(\frac{1}{2}\)
\[(\frac{1}{2})! =(\frac{3}{2}-1)! = \Gamma(\frac{3}{2}) = \Gamma(\frac{1}{2}+1) =\frac{1}{2}\Gamma(\frac{1}{2}) \]
With the integral I guess we can use \[\int\limits_{0}^{\infty} x^n e^{-ax} dx = \frac{ n! }{ a^{n+1} }\]
got that link through bibby
yeah but that is just begging the question because your numerator still has a factorial in it
Ooh ok I see what you did ganeshie
Thanks :P
Above works only when you can reduce a fraction down to \(\frac{1}{2}\) using the recurrence relation
Yup I see, \[\Gamma (1/2) = \sqrt{\pi}\] is the only one I see in my formulae sheet, so is that the only one defined?
For example, you can't find \(\Gamma(1/3)\) using just the above two goodies
Yeah you can find the gamma value of all half integers using above recurrence relation
Ah gotcha, good stuff! Thank you
See if you can you find a formula for \(\Gamma(\frac{n}{2})\)
\(n\in \mathbb{Z^+}\)
That seems tricky, I'll try
\[\Gamma(\frac{n}{2}) =(\frac{n}{2}-1)\Gamma(\frac{n}{2}-1) \]
Yeah haha, I was just going to ask you that, if I could start with \[\Gamma(\frac{ n }{ 2 }) =(\frac{ n }{ 2 }-1)!\] sort of deal
We use the symbol \(!\) only for integers use \(\Gamma\) for reals
Ah ok, I've never used gamma functions before I don't think until QM
I actually needed to know how to use this for here haha |dw:1453005957021:dw|
Let \(ax^2 = u \\\implies 2axdx = du \\\implies dx =\dfrac{du}{2a \sqrt{u/a}} \)
\[\color{#000000 }{ \displaystyle \Gamma (x)=\int\limits\limits_{0}^{\infty}z^{x-1}e^{-z}~dz } \]
Nice!
the integral then becomes \[\int\limits_0^{\infty} (\sqrt{u/a})^me^{-u} \dfrac{du}{2a\sqrt{u/a}}\\~\\=\frac{1}{2a^{(m+1)/2}}\int\limits_0^{\infty} u^{(m-1)/2}e^{-u}du\\~\\ =\frac{1}{2a^{(m+1)/2}}\int\limits_0^{\infty} u^{\color{red}{(m+1)/2}-1}e^{-u}du\\~\\ \] Compare that integeral with the gamma integral
Haha, now I know how they got that formula, that seems super simple..in hindsight
are you also working with beta function ?
there is a nice relation between gamma and beta functions
\[\beta(a,b) = \dfrac{\Gamma(a)\Gamma(b)}{\Gamma(a+b)}\]
Nope, we're really not even taught these gamma functions, but rather given a formula sheet with bunch of integrals on them...that's physics for you! Formulas, no need to prove
That is not physics, that is more like engineering. Physicists do care for proofs. Only dumb engineers look for formulas w/o proofs ;)
Haha, well yeah I guess I should've clarified, it's not that there is no need to prove, I suppose they assume you already know the proof at least in quantum mechanics!
It's weird...like these guassian integrals, it's so different from classical mechanics, all about probability
These special functions are not so hard. They are defined using definite integrals and if you're in QM, the definition is all you care for, I guess..
I still like the proofs though, but these are handy only because otherwise each problem will take like 10 years to do without them
Like check this integral out, what in the world |dw:1453007568253:dw|
I don't like to mindlessly use these integrals, but sometimes I think it's pretty important haha.
That integral is easy if you're familiar with the beta function
But I feel that you wont learn anything by evaluating that integral on your own. It's just algebra and you will get tired for sure. Better to use the result directly :)
Join our real-time social learning platform and learn together with your friends!