home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: sci.math
- Path: sparky!uunet!usc!sdd.hp.com!ux1.cso.uiuc.edu!news.cso.uiuc.edu!dbradley
- From: dbradley@symcom.math.uiuc.edu (David Bradley)
- Subject: Fractional Integral Problem
- Message-ID: <Bs7wup.IAp@news.cso.uiuc.edu>
- Sender: usenet@news.cso.uiuc.edu (Net Noise owner)
- Organization: Math Dept., University of Illinois at Urbana/Champaign
- Date: Thu, 30 Jul 1992 19:59:12 GMT
- Lines: 36
-
- Can anyone suggest some techniques for dealing with the following
- function involving a fractional integral?
- oo
- / -u b-1 -a
- Let F(a,b,z) = GAMMA(a) | e u (1+zu) du, Re(a)>0, Re(b)>0,
- /
- 0
-
- and complex z not on the negative real axis. I have reason to believe
- the above is symmetric in a and b, but have no rigorous proof.
-
- It's easy to show that the z-derivatives of S(a,b,z) := F(a,b,z)-F(b,a,z)
- all vanish at 0, but it appears F(a,b,z) is not analytic at z=0, so I
- can't equate S(a,b,z) with its Maclaurin series.
-
- The transformations v=zt, z=1/w give
-
- oo y
- a+b / -wy / a-1 b-1 -a
- F(a,b,z) = w | e | (y-v) v (1+v) dv dy,
- / /
- 0 0
-
- so it would be enough to show the inner integral is symmetric in a and
- b for each y>0. I can do this for 0<y<1 by expanding (1+v)^(-a), so
- that the inner integral becomes an infinite sum of Beta functions
- multiplied by suitable Gamma functions to make everything symmetric in
- a and b. Unfortunately, this idea doesn't seem to carry over when y>1.
-
- Any suggestions would be appreciated. I suppose I should add that this is
- not an assigned problem except insofar as I assigned it to myself. The
- integral arises in a proof for a continued fraction expansion for the
- incomplete Gamma function. Reference: H.S. Wall, Analytic Theory of
- Continued Fractions.
-
- -D.M.Bradley
-