Hi everybody,
I am again stuck on an optimization problem: I have a function of 6 variables that I want to optimize. I want:
0 < x(1) < 24
0 < x(2) < 24
24 < x(3) < 48
0 < x(4) < 24
48 < x(5) < 72
0 < x(6) < 24
A precision of 0.1 would be more than enough for those 6 parameters! I used:
Results = fmincon(@(x)optforecastSOC72(x,aConst),x0,[],[],[],[],[0,0,24,0,48,0],[24,24,48,24,72,24],[],options)
And I get a result which is NOT the minimum over that interval. How can I force the fmincon to go the entire interval (maybe with bigger sizestep)? I tried to change every options in optimset but it didn't change a think ;(. Thanks for your advices

 Respuesta aceptada

Alan Weiss
Alan Weiss el 11 de Jun. de 2014

1 voto

It is not clear in what way the result does not satisfy you. If you think fmincon did not arrive at a local minimum, then perhaps your objective function is not smooth. Consult Optimizing a Simulation or ODE for suggestions.
If, however, your objective function is smooth, but fmincon returns a local minimum that is not a global minimum, then consult Local Vs. Global Optima.
Good luck,
Alan Weiss
MATLAB mathematical toolbox documentation

Más respuestas (0)

Etiquetas

Preguntada:

el 10 de Jun. de 2014

Respondida:

el 11 de Jun. de 2014

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by