MathGroup Archive 1997

[Date Index] [Thread Index] [Author Index]

Search the Archive

Re: trouble integrating Bessel Functions

  • To: mathgroup at smc.vnet.net
  • Subject: [mg6500] Re: [mg6484] trouble integrating Bessel Functions
  • From: jpk at max.mpae.gwdg.de
  • Date: Thu, 27 Mar 1997 02:42:36 -0500 (EST)
  • Sender: owner-wri-mathgroup at wolfram.com

> From pankratz at eta.pha.jhu.edu Tue Mar 25 05:50:30 1997
> Date: Mon, 24 Mar 1997 21:38:31 -0500 (EST)
> From: Chris Pankratz <pankratz at eta.pha.jhu.edu>
> To: mathgroup at smc.vnet.net
> Subject: [mg6484] trouble integrating Bessel Functions
> 
> I have encountered a curious problem calculating the fourier transform of a
> modified bessel function.  Mathematica 2.2 will calculate it just fine, but
> Mathematica 3.0 won't.  Can anyone assist?  Thanks.
> Chris
> 
> Mathematica 2.2 does it just fine:
> ----------------------------------
> 
> (1/Sqrt[2 Pi]) * Integrate[BesselK[0, t] Exp[I w t], {t, -Infinity,
> Infinity}]
> 
>        Pi
>   Sqrt[--]
>        2
> ------------
>           2
> Sqrt[1 + w ]
> 
> 
> Mathematica 3.0 doesn't seem to have the right integration package loaded
> -------------------------------------------------------------------------
>    - it just echoes the command back to me, without giving a result.
> 
> 
Oh,  Your first integral is Infinity and Mma 2.2 doesn't report this. Because
You don't want a Besselfunction wit a negative argument. The imaginary part
of K[0,t] diverges for t<0 !
When You give

Integrate[BesselK[0,t]*Exp[I*w*t],{t,0,Infinity},
          Assumptions->Im[w]==0]/Sqrt[2 Pi] 

You get the correct result.

Hope that helps
Jens
My Abramowitz/Stegun say
Integrate[ t^mu K[mu,t],{t,0,Infinity}==2


  • Prev by Date: solving simple ODE using NDSolve
  • Next by Date: Symbols, names, objects
  • Previous by thread: trouble integrating Bessel Functions
  • Next by thread: How can I handle Operator Algebra ?