Results 1 to 5 of 5

Math Help - [SOLVED] Incomplete Gamma Function

  1. #1
    Super Member Deadstar's Avatar
    Joined
    Oct 2007
    Posts
    722

    [SOLVED] Incomplete Gamma Function

    I already posted this in the Applied section but I don't think thats actually the right place so here it is again!

    I think I'm pretty close to getting this but can't quite figure out what do next!

    From earlier in the question...

    \textrm{erf}(x) = \frac{2}{\sqrt{\pi}} \int^x_0 e^{-t^2} dt \sim 1 - \frac{e^{-x^2}}{\pi} \sum^{\infty}_{n=0} \Gamma \Big{(}n + \frac{1}{2}\Big{)}\frac{(-1)^n}{x^{2n+1}} (*)

    Show that erf(x) can be expressed in terms of the incomplete gamma function, and that the above asymptotic expansion agrees with the result obtained in Problem 6. [The formula \Gamma\Big{(}\frac{1}{2} + n)\Gamma\Big{(}\frac{1}{2} - n\Big{)} = (-1)^n \pi (**) for integer n may be useful.]

    So I think I have to show the relation... \gamma\Big{(}\frac{1}{2}, x\Big{)} = \sqrt{\pi} \textrm{erf}(\sqrt{\pi}) (***) as that's the relation I've seen elsewhere.

    The result obtained in question 6 was...

    \gamma(\alpha,x) \sim \Gamma(\alpha) - e^{-x}x^{\alpha - 1} \sum^{\infty}_{n=0}\frac{\Gamma(\alpha)}{\Gamma(\a  lpha - n)}\frac{1}{x^n}

    And also...
    \gamma(\alpha,x) \sim \Gamma(\alpha) - e^{-x}x^{\alpha - 1} \sum^{\infty}_{n=0}\frac{(\alpha-1)(\alpha-2)\ldots(\alpha-n)}{x^n}

    Right so intro done!

    Now what I've done...

    From (*)
    \sqrt{\pi} \textrm{erf}(x) = 2 \int^x_0 e^{-t^2} dt \sim \sqrt{\pi} - \frac{e^{-x^2}}{\sqrt{\pi}} \sum^{\infty}_{n=0} \Gamma\Big{(}n + \frac{1}{2}\Big{)}\frac{(-1)^n}{x^{2n+1}}

    = \Gamma\Big{(}\frac{1}{2}\Big{)} - \frac{e^{-x^2}}{\sqrt{\pi}} \sum^{\infty}_{n=0} \Gamma\Big{(}n + \frac{1}{2}\Big{)}\frac{(-1)^n}{x^{2n+1}}
    (This step seems to be showing that \alpha = \frac{1}{2} which corresponds to (***))

    Using (**)
    = \Gamma\Big{(}\frac{1}{2}\Big{)} - \sqrt{\pi} e^{-x^2} \sum^{\infty}_{n=0} \frac{1}{\Gamma\Big{(}\frac{1}{2} - n\Big{)}}\frac{1}{x^{2n+1}}

    = \Gamma\Big{(}\frac{1}{2}\Big{)} - \sqrt{\pi} e^{-x^2} \sum^{\infty}_{n=0} \frac{x^{-(n+1)}}{\Gamma\Big{(}\frac{1}{2} - n\Big{)}}\frac{1}{x^n}

    = \Gamma\Big{(}\frac{1}{2}\Big{)} - e^{-x^2} \sum^{\infty}_{n=0} \frac{\Gamma\Big{(}\frac{1}{2}\Big{)}}{\Gamma\Big{  (}\frac{1}{2} - n\Big{)}}\frac{x^{-(n+1)}}{x^n}


    But I'm stuck now. I need to do something to the e^{-x^2} \textrm{ and the } x^{-(n+1)} but I'm not sure what... I don't even know if what I've done so far is right to be honest... Any help please!
    Last edited by Deadstar; February 1st 2010 at 03:37 PM. Reason: removed sqrt(pi) from last equation
    Follow Math Help Forum on Facebook and Google+

  2. #2
    MHF Contributor Drexel28's Avatar
    Joined
    Nov 2009
    From
    Berkeley, California
    Posts
    4,563
    Thanks
    21
    Quote Originally Posted by Deadstar View Post
    I already posted this in the Applied section but I don't think thats actually the right place so here it is again!

    I think I'm pretty close to getting this but can't quite figure out what do next!

    From earlier in the question...

    \textrm{erf}(x) = \frac{2}{\sqrt{\pi}} \int^x_0 e^{-t^2} dt \sim 1 - \frac{e^{-x^2}}{\pi} \sum^{\infty}_{n=0} \Gamma \Big{(}n + \frac{1}{2}\Big{)}\frac{(-1)^n}{x^{2n+1}} (*)

    Show that erf(x) can be expressed in terms of the incomplete gamma function, and that the above asymptotic expansion agrees with the result obtained in Problem 6. [The formula \Gamma\Big{(}\frac{1}{2} + n)\Gamma\Big{(}\frac{1}{2} - n\Big{)} = (-1)^n \pi (**) for integer n may be useful.]

    So I think I have to show the relation... \gamma\Big{(}\frac{1}{2}, x\Big{)} = \sqrt{\pi} \textrm{erf}(\sqrt{\pi}) (***) as that's the relation I've seen elsewhere.

    The result obtained in question 6 was...

    \gamma(\alpha,x) \sim \Gamma(\alpha) - e^{-x}x^{\alpha - 1} \sum^{\infty}_{n=0}\frac{\Gamma(\alpha)}{\Gamma(\a  lpha - n)}\frac{1}{x^n}

    And also...
    \gamma(\alpha,x) \sim \Gamma(\alpha) - e^{-x}x^{\alpha - 1} \sum^{\infty}_{n=0}\frac{(\alpha-1)(\alpha-2)\ldots(\alpha-n)}{x^n}

    Right so intro done!

    Now what I've done...

    From (*)
    \sqrt{\pi} \textrm{erf}(x) = 2 \int^x_0 e^{-t^2} dt \sim \sqrt{\pi} - \frac{e^{-x^2}}{\sqrt{\pi}} \sum^{\infty}_{n=0} \Gamma\Big{(}n + \frac{1}{2}\Big{)}\frac{(-1)^n}{x^{2n+1}}

    = \Gamma\Big{(}\frac{1}{2}\Big{)} - \frac{e^{-x^2}}{\sqrt{\pi}} \sum^{\infty}_{n=0} \Gamma\Big{(}n + \frac{1}{2}\Big{)}\frac{(-1)^n}{x^{2n+1}}
    (This step seems to be showing that \alpha = \frac{1}{2} which corresponds to (***))

    Using (**)
    = \Gamma\Big{(}\frac{1}{2}\Big{)} - \sqrt{\pi} e^{-x^2} \sum^{\infty}_{n=0} \frac{1}{\Gamma\Big{(}\frac{1}{2} - n\Big{)}}\frac{1}{x^{2n+1}}

    = \Gamma\Big{(}\frac{1}{2}\Big{)} - \sqrt{\pi} e^{-x^2} \sum^{\infty}_{n=0} \frac{x^{-(n+1)}}{\Gamma\Big{(}\frac{1}{2} - n\Big{)}}\frac{1}{x^n}

    = \Gamma\Big{(}\frac{1}{2}\Big{)} - \sqrt{\pi} e^{-x^2} \sum^{\infty}_{n=0} \frac{\Gamma\Big{(}\frac{1}{2}\Big{)}}{\Gamma\Big{  (}\frac{1}{2} - n\Big{)}}\frac{x^{-(n+1)}}{x^n}


    But I'm stuck now. I need to do something to the e^{-x^2} \textrm{ and the } x^{-(n+1)} but I'm not sure what... I don't even know if what I've done so far is right to be honest... Any help please!
    \Gamma\left(\frac{1}{2},x\right)=\int_0^x z^{\frac{-1}{2}}e^{-z}\text{ }dz, let \sqrt{z}=\xi\implies z=\xi^2\implies dz=2\xi\text{ }d\xi. So, we see that \Gamma\left(\frac{1}{2},x\right)=\int_0^{\sqrt{x}}  \frac{e^{-\xi^2}}{\xi}\cdot 2\xi\text{ }d\xi=2\int_0^{\sqrt{x}}e^{-\xi^2}\text{ }d\xi=\sqrt{\pi}\text{erf}\left(\sqrt{x}\right).
    Follow Math Help Forum on Facebook and Google+

  3. #3
    Super Member Deadstar's Avatar
    Joined
    Oct 2007
    Posts
    722
    Thanks for that!

    Was my way actually going anywhere or not really..?

    While your proof is good...
    Does it not require the knowledge to use \alpha = \frac{1}{2}, and doesn't use the \Gamma(\frac{1}{2} + n) \Gamma(\frac{1}{2} - n) = (-1)^n \pi hint which makes me think it is supposed to be solved some other way...

    Or is what I've done more to do with the 'and that the above asymptotic expansion agrees with the result obtained in Problem 6.' part?
    Follow Math Help Forum on Facebook and Google+

  4. #4
    MHF Contributor Drexel28's Avatar
    Joined
    Nov 2009
    From
    Berkeley, California
    Posts
    4,563
    Thanks
    21
    Quote Originally Posted by Deadstar View Post
    Thanks for that!

    Was my way actually going anywhere or not really..?

    While your proof is good...
    Does it not require the knowledge to use \alpha = \frac{1}{2}, and doesn't use the \Gamma(\frac{1}{2} + n) \Gamma(\frac{1}{2} - n) = (-1)^n \pi hint which makes me think it is supposed to be solved some other way...

    Or is what I've done more to do with the 'and that the above asymptotic expansion agrees with the result obtained in Problem 6.' part?
    There is a lack of context here. I am not using your book and so I do not know what it requires or what it needs. Honestly, I just saw that you said to prove that \Gamma\left(\frac{1}{2},x\right)=\sqrt{\pi}\text{e  rf}\left(\sqrt{x}\right).

    I'll come back later and read it more thoroughly.
    Follow Math Help Forum on Facebook and Google+

  5. #5
    Super Member Deadstar's Avatar
    Joined
    Oct 2007
    Posts
    722
    Ah I see!

    That's just the relationship I found elsewhere so I had an idea of how to approach this, it's not actually mentioned in the question.

    Basically, this is question 7. Question 6 was to find a series expansion of the incomplete gamma function which was shown to be...

    Quote Originally Posted by Deadstar View Post
    \gamma(\alpha,x) \sim \Gamma(\alpha) - e^{-x}x^{\alpha - 1} \sum^{\infty}_{n=0}\frac{\Gamma(\alpha)}{\Gamma(\a  lpha - n)}\frac{1}{x^n}
    Then for question 7 you had to show that the error function could be written as...

    Quote Originally Posted by Deadstar View Post
    \textrm{erf}(x) = \frac{2}{\sqrt{\pi}} \int^x_0 e^{-t^2} dt \sim 1 - \frac{e^{-x^2}}{\pi} \sum^{\infty}_{n=0} \Gamma \Big{(}n + \frac{1}{2}\Big{)}\frac{(-1)^n}{x^{2n+1}} (*)
    Which I've also done.

    Finally, the part that I'm stuck on is...
    Quote Originally Posted by Deadstar View Post
    Show that erf(x) can be expressed in terms of the incomplete gamma function, and that the above asymptotic expansion agrees with the result obtained in Problem 6. [The formula \Gamma\Big{(}\frac{1}{2} + n)\Gamma\Big{(}\frac{1}{2} - n\Big{)} = (-1)^n \pi (**) for integer n may be useful.]
    And I basically manipulated the series expansion of erf(x) to try and get close to the series expansion of the incomplete gamma function. But I've hit a brick wall at the last line in my original post. If you need more info I could just PM you a jpg of the question but I hope you can figure out what I'm asking!
    Follow Math Help Forum on Facebook and Google+

Similar Math Help Forum Discussions

  1. Incomplete gamma function
    Posted in the Differential Geometry Forum
    Replies: 10
    Last Post: May 31st 2011, 10:28 PM
  2. [SOLVED] The error function and incomplete gamma function
    Posted in the Advanced Applied Math Forum
    Replies: 0
    Last Post: January 31st 2010, 03:24 PM
  3. [SOLVED] using beta or gamma function
    Posted in the Calculus Forum
    Replies: 5
    Last Post: January 5th 2010, 11:02 PM
  4. [SOLVED] Gamma function simple help needed
    Posted in the Calculus Forum
    Replies: 9
    Last Post: June 23rd 2008, 01:46 PM
  5. Incomplete Graph of a Rational Function
    Posted in the Algebra Forum
    Replies: 1
    Last Post: May 21st 2008, 09:15 PM

Search Tags


/mathhelpforum @mathhelpforum