From help-octave-request at octave dot org Mon Apr 26 12:19:24 2004
Subject: Re: backward ODE solving
From: "Ben Diedrich" <Ben dot Diedrich at noaa dot gov>
To: "John W. Eaton" <jwe at bevo dot che dot wisc dot edu>
CC: help at octave dot org
Date: Mon, 26 Apr 2004 13:19:03 -0400

The ODE solver is called within a gradient optimization function that can
see a wide variety of problems. The current configuration uses the timesteps
selected by 'ode23' (either Matlab or octave-forge). If lsode could output
the internally selected timesteps, the behavior would be the same. I'm not
certain what impact pre-selected timesteps would have on the behavior of the
algorithm for different problems. I can try it out, though.

Ben

"John W. Eaton" wrote:

> On 26-Apr-2004, Ben Diedrich <Ben dot Diedrich at noaa dot gov> wrote:
>
> | I haven't tried 'lsode' yet, but I'd prefer to use a solver
> | that calculates the output time vector, instead of requiring one ahead
> | of time as lsode does.
>
> Why?  The times you specify for lsode are the times for which you want
> to see the solution.  Internally, it selects whatever timesteps are
> necessary to satisfy the local error tolerances that you specify for
> the solution.
>
> jwe
>
> -------------------------------------------------------------
> Octave is freely available under the terms of the GNU GPL.
>
> Octave's home on the web:  http://www.octave.org
> How to fund new projects:  http://www.octave.org/funding.html
> Subscription information:  http://www.octave.org/archive.html
> -------------------------------------------------------------



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------