[R-sig-dyn-mod] Time steps of integration

Thomas Petzoldt Thomas.Petzoldt at tu-dresden.de
Fri Nov 15 18:08:45 CET 2013


Hi,

the solver interpolates *during* the integration. Saving all
intermediate steps would be much too inefficient and memory consuming.

Sometimes, considerable work is needed to identify a good automatic
stepping and sometimes, steps are discarded and the solver goes a little
bit back.

In case of events, time steps are calculated this way, that events are
matched. If no events are present, integration is allowed to overrun the
external steps.

If you have external forcings, step size should be limited with hmax.


Thomas



More information about the R-sig-dynamic-models mailing list