I have a simulink model with some parameters and i want to find the minimum of an objective function. The problem is that i should run the minimization for many times varying the objective function itself, but with normal settings it takes me 7,5 min to get only one estimation, too much if i need 100 of them. But the good thing is that i don't need a good precision in finding the minimum, some variables could be integer but fmincon doesn't allow to specify that. So i decided to relax some tolerances, in particolar StepTolerance. The problem is that at the start of simulation the step size is very small and this force me to set a bigger StepTolerance than needed. So i thought that maybe there was a method to impose a minimum step size and so i found FiniteDifferenceStepSize and DiffMinChange but in any way i set them, the inizial step size is in the order of e-3 but i would like it to be 1 or even bigger so that i can set StepTolerance to 1.
Regardless of what i wrote until now, can you suggest me a way to speed up my estimation so that i can get close to a minimun with a precision that i can choose? Maybe i'm not using correctly the previous options or maybe there is a better way to do the entire thing. Thanks in advance,
Federico Spada
