Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
lsqnonlin Levenberg-Marquard
148 views
Skip to first unread message
piera
Apr 16, 2013, 12:07:08 PM4/16/13
to
dear all: when executing lsqnonlin with the default 'trust-region-reflective' algorithm, matlab for some runs prints the message:
"Warning: Trust-region-reflective algorithm requires at least as many equations as variables; using Levenberg-Marquardt algorithm instead."
Is my interpretation correct that in these cases it automatically switches to Levenberg-Marquardt algorithm (without me having to change it in the options) while for the other cases it uses the default algorithm?
thanks in advance.
"Warning: Trust-region-reflective algorithm requires at least as many equations as variables; using Levenberg-Marquardt algorithm instead."
Is my interpretation correct that in these cases it automatically switches to Levenberg-Marquardt algorithm (without me having to change it in the options) while for the other cases it uses the default algorithm?
thanks in advance.
Alan_Weiss
Apr 16, 2013, 2:20:37 PM4/16/13
to
Alan Weiss
MATLAB mathematical toolbox documentation
0 new messages