Re: [Help-gsl] integration

[mail]

change of variables needs the integrand to be invertible
over the interval - which is not the case, try it the other
way round (shorten interval to [r,1] and regard periodics)
you should get (1-2*cos(a)+cos(a)^2)/a^2 with a = 10^8.

On 15 May 2006 at 9:49, ghadir mohammadkhani wrote:

> Dear My friends
>  I would like to thank you for GNU Scientific Library.
> I have a problem in respect to oscillatory integrals.
> For example, I want to Integrate from a very simple
> function that its solution is perfectly evident
> analytically:
> Sin(100,000,000*x)*Sin(100,000,000*y) where x and y go
> from 0-1. I have done a variable change over x and y,
> with this work, x and y go from 0-10^8, since the
> sinuous integral at interval 0-2*pi is zero, so I
> can choose integral interval as following:
> 10^8/(2*pi)-Integer part [10^8/(2*pi)]=0.30919
> I have used several subroutines, but unfortunately I
> have not obtained an exact result. Now, I wanted to
> know whether there is a special way for
> these kinds of integrals?
> If it is possible for you, Please guide me.
> Thank you so much all of frieds for your kind
> attention and consideration.
> With best regards,
> Ghadir Mohammadkhani.
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around 
> http://mail.yahoo.com 
> 
> 
> 
> _______________________________________________
> Help-gsl mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/help-gsl