[R] evolution of Nelder-Mead process

kostas savvidis ksavvidis at gmail.com
Tue Feb 9 13:02:25 CET 2010

The optim function has a debug option called trace see the helppage:

Non-negative integer. If positive, tracing information on the progress
of the optimization is produced. Higher values may produce more
tracing information: for method "L-BFGS-B" there are six levels of
tracing. (To understand exactly what these do see the source code:
higher levels give more detail.)

for Nelder-Mead this outputs quite a bit of information. You would
only need more if you were debugging optim, as a user if you think
visualizing how the damned thing walks will help you understand your
problem better, you are probably onto the wrong path.

That said, you seem to have data which is defined on a grid, the
appropriate method in this case in SANN, not Nelder-Mead.

Hope this helps,


Nanjing University

+8625 8622 8040 (h)
+86 13451911944 (m)

More information about the R-help mailing list