non linear system of equations

Hi, I am trying to solve a non linear system (6x6) for different parameter values. I have mainly used the trust region dogleg method in the optimtool, which seems to find solutions but (1) either the system is solved but the solution is not a "real" solution (2) or there are inaccuracies which are reflected in huge errors when I perform some checks.

I don't know whether it can help, but rewriting the first two equations has proved to be useful as I don't get crazy solutions any more. However, they are too close to zero, I believe.

Considering that computing analytical derivatives is out of question, what would you suggest me to do?

Thanks in advance for your help.

Answer 1

[Alan Weiss]

I don't really understand what you are asking, but perhaps some documentation might help you figure out what to try next. You might also want to look in other areas of the Improving Results documentation.

Good luck,

 Alan Weiss
 MATLAB mathematical toolbox documentation

[Sara]

Dear Alan,

thank you for your reply. I have already tried to change parameters, initial points, relax tolerances but nothing seems to work. I always get convergence, but to a point which is not a zero of the system.

What I am asking is whether I should try other algorithms than the ones available in optimtool box (which and how?) or just keep going with fsolve.

Thank you in advance,

Sara