Dear Matthias, thanks a lot for your work. However, I have implemented a numeric workaround that does not need symbolic integration a few weeks ago. Since this approach works and I am running out of time with my PhD, I will unfortunately not be able to look further into ginac. At some point in the future maybe. Best regards and thanks again, Matthias Am Mon, 27 Apr 2015 16:27:43 +0200 schrieb Matthias Dellweg <dellweg@tp1.uni-duesseldorf.de>:
Am Mon, 27 Apr 2015 11:37:39 +0100 schrieb "Vladimir V. Kisil" <kisilv@maths.leeds.ac.uk>:
Dear All,
What is easy to do is to add to function class a primitive feature analogously to their knowledge of derivatives. Then, integral::eval_integ() may call such methods to evaluate simple integrals of plain functions.
If this looks reasonable to be included in GiNaC, then I may write such a patch.
Best wishes, Vladimir
> On Sat, 14 Mar 2015 17:35:10 +0100, Matthias Reis > <matthias.reis@biologie.hu-berlin.de> said:
MR> Dear all, as far as I understand, ginac cannot integrate MR> anything other than polynomials. How hard would it be (for MR> me maybe) to implement exponentials? What needs to be done? MR> This feature would be great for me. I just need to MR> integrate simple e^{x * constant} terms.
MR> Best regards, Matthias
Dear Matthias, i have an implementation ready that integrates expressions like step(x-a)*x^5*exp(b*x). You can find it at https://bitbucket.org/mdellweg/ginac.git . It just can not be done in the way described above, since it needs to know stuff outside of exp(...). I would love to hear comments on this. Kindest regards, Matthias