[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: IVP for Parabolic-Elliptic 1D Queries
From: |
c. |
Subject: |
Re: IVP for Parabolic-Elliptic 1D Queries |
Date: |
Fri, 7 Mar 2014 17:59:24 +0100 |
On 7 Mar 2014, at 15:38, prao <address@hidden> wrote:
> fgnievinski wrote
>> Any insight from how Scilab, Freemat, Julia, SciPy, or R handle this?
>> -F.
>
> Hi Felipe,
>
> Thanks for the reply. Different packages have different ways to handle it,
> and the methods have their own pros and cons. I compiled a list of some of
> the packages including the ones you suggested:
>
> 1. R
> Uses finite difference (mainly a combination of Runge-Kutta and multistep
> methods like BDF and Adams)
>
> 2. Netlib
> It has subroutines employing different methods
> -PDECOL (B-splines collocation)
> -EPDCOL (B- splines)
> -PDEONE ( finite difference)
> -PDECHEB (C0 collocation)
>
> 3. Julia
> It doesn't have a PDE solver but it does have an ODE solver that interfaces
> with Sundials(finite difference, multi step methods)
>
> 4. DUNE
> It has a bunch of solvers. Apparently their hybrid Galerkin has very good
> performance.
>
> 5. SciPy doesn't have a PDE solver but FiPy(finite volume) and StePy(finite
> element) do.
>
> 6. Freemat
> Couldn't find anything related to pde solvers
>
> I am leaning towards finite difference methods using Runge-Kutta type
> methods. My preference is based on my familiarity and relative use of their
> implementation and their high performance. I am doing some more paper
> reading to figure out exactly what RK methods would be appropriate.
>
> Does anyone have any ideas, comments or suggestions? Thanks!
>
Octave is missing in this list.
Do you know what are currently the options for solving PDEs available in Octave?
> Best,
> Pooja
c.