| IDL question about IMSL_NLINLSQ [message #67461] |
Wed, 22 July 2009 02:21 |
wy
Messages: 2
Registered: July 2009
|
Junior Member |
|
|
I'm using IMSL_NLINLSQ to solve nonlinear equations right now. I've
got results utilizing argument like "MAX_STEP=0.01,TOL_AFCN=0.01", but
I have an prompt as below:
"IMSL Error: IMSL_NLINLSQ: Warning: MATH_UNBOUNDED: Five consecutive
steps of length "MAX_STEP" have been taken; either the function is
unbounded below, or has a finite asymptote in some direction or the
maximum allowable step size "MAX_STEP" is too small."
I changed the arguments to be more sensitive seemingly like
"MAX_STEP=0.0001,TOL_AFCN=0.00001", but the computer ran much quicker.
It's rather terrible. (The prompt also existed.)
In fact, I can get entirely different results from MATLAB 'fsolve'
function using the same Levenberg-Marquardt algorithm. The results
from MATLAB don't change with initial values and the residuals are
negligible, while residuals from IDL lie in between 0.1 and 0.2.
I'm really confused. Could someone help me explain this?
|
|
|
|
comp.lang.idl-pvwave archive
Members
Search
Help
Login
Home
