[R-sig-dyn-mod] Problem with simecol and ode - crash from R
dave fournier
otter at otter-rsch.com
Sat Nov 28 18:10:03 CET 2009
Artificially bounding parameters reduces the differentiability
of the function and makes optimization slower or more difficult.
a better approach I think is to use implicit methods and fixed step size
for formulating the solution to the ODE. See section 1-17 of the ADMB
manual for this approach applied to the Schaeffer-Pella-Tomlinson model.
http://admb-project.googlecode.com/files/admb.pdf
That is the name used in fisheries for a generalization of the
logistic. Using these ideas and ADMB it is possible to quickly fit
multi-parameter generalizations of these models.
The big advantages of the implict method are that it is numerically
stable and can use a larger step size.
--
David A. Fournier
P.O. Box 2040,
Sidney, B.C. V8l 3S3
Canada
Phone/FAX 250-655-3364
http://otter-rsch.com
More information about the R-sig-dynamic-models
mailing list