[R-sig-dyn-mod] a question on ODE.1D, nspec

[Thomas Petzoldt]

On 6/26/2013 10:47 PM, Dabing Chen wrote:
> Hi Daniel: I just tried ODE, it worked. Is it true that we do not
> want to use ODE.1D if the length is not n*N? I am also wondering what
> is the benefit of ODE.1D over ODE? It seems that we can use ODE to
> finish the job. Thanks, Dabing

Some of the solvers used by ode.1D ... ode.3D make use of the regular 
structure of such problems specified by nspec resp. dimens, so that an 
adequate Jacobian matrix can be constructed instead of using the full 
Jacobian. This makes simulations faster, especially for large systems.

See ?lsodes for more details.

Another advantage are the specific plotting functions for results of ode.1D.


Thomas