fsolve returns non-integer exitflag

Hi everyone,

I am experiencing a strange problem. For as far as I know an exitflag should always be an integer (identifying the reason the algorithm terminated). However when solving my problem with fsolve, the exitflag turns out to be a non-integer: 1.329410886565489.

fsolve did not find a solution, but that is perfectly possible. The exitflag being a non-integer value on the other hand seems really strange to me.

What could be the problem?

Kind regards, Corrado Sirianni

[Mischa Kim]

Just out of curiosity, could you attach your code?

[Corrado]

I would like to, but that's somewhat complicated because it's a really long script (divided over several functions and m-files) leading to the equations to be solved (system of two equations). I'm computing the admissibility domain for expansion shock waves in the near-critical region of methane, carbon dioxide and ethylene.

[Corrado]

Actually, I have the same problem with any other equations I'm trying to solve with fsolve. The exitflag is never an integer...

[Matt J]

Can you at least show your call to FSOLVE, and maybe a bit of the surrounding code like the set-up of the initial x0 and the selection of options using OPTIMSET or OPTIMOPTIONS?

[Corrado]

for iso_GAMMA = 2

dlmwrite('Prho_coordinates_iso_gamma.dat',[iso_GAMMA iso_GAMMA],'delimiter',' ','newline','pc','-append')

for teta = -1:0.01:1

[r,exitflag] = fsolve(@fun_iso_gamma,0.01)

if r > 0.0585

    continue

end

if r < 0

    continue

end

. . . . . . etc. etc.

Answer 1

[Alan Weiss]

I guess that you are interpreting the output of fsolve incorrectly. As stated in the function reference page, the outputs of fsolve are, in this order:

 [x,fval,exitflag,output,jacobian] = fsolve(...)

exitflag is always the third output. I would bet that the third output that you obtain from solve is an integer, whatever name you give it.

Alan Weiss

MATLAB mathematical toolbox documentation

[Corrado]

You are totally right! I forgot the order is also of importance! Thank you very much Alan!