help-octave
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: bvp4c not implemented in Octave 4.0?


From: Nicholas Jankowski
Subject: Re: bvp4c not implemented in Octave 4.0?
Date: Thu, 16 Jul 2015 11:35:18 -0400

On Thu, Jul 16, 2015 at 4:50 AM, Carlo De Falco <address@hidden> wrote:

On 16 Jul 2015, at 09:36, Tatsuro MATSUOKA <address@hidden> wrote:

> As Calro wrote, if you need efficiency, it is necessary to polish the code.

What bvp4c currently does is just set up the system of nonlinear equations
for the discretized differential problem then let fsolve deal with the solution,
IIRC it doesn't even build a Jacobian. After error estimation the problem is
solved again from scratch on a new grid.

So the solution should be correct but the implementation approach is really not
efficient, the correct algorithm is described in full detail in Shampine's
book so if anyone is interested it should be quite straightforward to implement.

c.
_______________________________________________
Help-octave mailing list
address@hidden
https://lists.gnu.org/mailman/listinfo/help-octave


All that said, if the package is released with a functionally correct (even if slow) implementation, Octave should produce the warning:

warning: the 'bvp4c' function belongs to the odepkg package from Octave Forge
which you have installed but not loaded.

I don't know if there's some other appropriate way (perhaps in the help? or a warning when running?) to flash a warning that 'this will work for you, but don't expect much unless you come and help fix it'

reply via email to

[Prev in Thread] Current Thread [Next in Thread]