Fsolve and initial guesses

Hi,

I have a system of nonlinear equations and corresponding variabls that I need to solve for. I use the 'fsolve' function and I get the following message from MATLAB:

Equation solved, inaccuracy possible.

The vector of function values is near zero, as measured by the value

of the function tolerance. However, the last step was ineffective.

<stopping criteria details>

I noticed that the solutions change a bit - at least some variables qualitatively - once I change initial guesses. I keep getting the same message as above but qualitatively different solutions when I increase all initial guesses by 10% and again when I decrease all guesses by 10%. How worried should I be?

Answer 1

[Torsten]

You already received several responses for this question here:

https://uk.mathworks.com/matlabcentral/answers/1973704-interpreting-the-results-of-fsolve?s_tid=prof_contriblnk

Iterative processes keep the risk of converging to no solution or to a wrong solution. So the only way to decide "how worried you should be" is to look at the solution, see whether it makes physical sense and compute whether it gives a residual near zero when inserted into your system of equations.