fsolve

[John W. Eaton]

On 10-May-1999, Ursula M. Molter <address@hidden> wrote:

| I am new to Octave (i recently installed version 2.0.14)
| 
| I am trying tu run fsolve -
| I got very good results (precision of e-11) but when I increased
| the precision,
| I got the Info message: iteration is not converging well
| and fsolve did not change the values.
| Where does fsolve store the intermediate values?

Currently, Octave's fsolve function uses a MINPACK subroutine to do
the real work.  The C++ class that actually calls the Fortran
subroutine doesn't currently have a way to access or return any
intermediate values.

| is it possible to force it to output them?

Not without changing the C++ sources.  However, since you have the
source (or can get it), it is possible to fix the problem.

jwe