[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gsl] bug in GSL numerical integration
From: |
Pedro Gonnet |
Subject: |
Re: [Bug-gsl] bug in GSL numerical integration |
Date: |
Tue, 13 Aug 2013 16:23:20 +0100 |
Hi Nikhil,
In that case I would suggest using a more robust integrator, e.g.
gsl_integration_cquad (disclaimer: I wrote this integrator myself), and
doing the integral substitution yourself.
There are a number of different substitution functions that may be more
or less well adapted to your problem. A good place to start is [1] which
describes some different substitutions and their properties.
Cheers,
Pedro
[1] http://www.math.ethz.ch/~waldvoge/Projects/nisJoerg.pdf
On Tue, 2013-08-13 at 13:24 +0000, Nikhil wrote:
> Hi Pedro,
>
> Thank you very much for your response. I got your point and why my
> integration returns 0.0 (since the stretch/shrink factor in the
> transformation goes as 1/t^2, higher patches on t will be stretched more) :
> (.
>
> Increasing tolerance did not help in this case. I don't know what other
> method could work. I have an integral, which is spread over a large span
> along x-axis.
>
> regards,
> Nikhil
>
>