Re: [Help-gsl] Oscillating integral: Product of Bessel functions

[Mehmet Suzen]

Hi,
Actuall main function looks like

 f = int_{0}^{infty} x * cos(-x*Z) * gsl_sf_bessel_In(m,x*R) *
gsl_sf_bessel_Kn(m+1,x*R)
Z and R parameters. For this also I got problems with QAWO and QAWF but
if I use QAGS I can get a result but it is oscillating with increasing m.
Any suggestion ?

Thanks,
Mehmet
On 3/9/06, Inigo Aldazabal Mensa <address@hidden> wrote:
> Hi,
>
> As x->\infty your integrand diverges oscillating something like exponentially,
> as gsl_sf_bessel_In does, and nothing kills this behaviour. Are you sure this
> integral should converge?
>
> Regards,
>
> Iñigo
>
> El Jueves, 9 de Marzo de 2006 09:54, Mehmet Suzen escribió:
> > Hello All,
> > I have tried to use QAWO or QAWF for integrating function which is
> > a product of modified bessel functions. Integral looks like
> >  f = int_{0}^{infty} x * cos(-x*Z) * gsl_sf_bessel_In(m,x*R)  dx
> > X and R are parameters.
> > but I am getting  error message ;
> > gsl: qawo.c:436: ERROR: cannot reach tolerance because of roundoff error
> > I have tried different epsabs or TOL, it doesn't help. How do I choose
> > right tolerance and bin sizes? And I have seen on the net that product of
> > bessel functions are problematic in integration. Any comment would be
> > helpful. Cheers,
> > --Mehmet
> >
> >
> > _______________________________________________
> > Help-gsl mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/help-gsl
>
>
> _______________________________________________
> Help-gsl mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/help-gsl
>


--

--Mehmet Suzen
_________________________________________
Entia non sunt multiplicanda praeter necessitatem
_________________________________________