Optimizing simulink parameters from matlab code

I have a simulink simulation that I am trying to optimize, however the gains (gain a,b,c etc) are calculated outside of simulink using matlab code before running the simulation.

For example if you want to run the simulation, you first go to the matlab code to calculate all the parameters:

In the matlab script one might enter:

Q=12000

f=24000

and from that the gains are calculated in the matlab script to be:

a=1/Q

b=f*Q

etc

I would like to optimize the values of Q and f rather than the gains a,b,c. However, when I optimize, my only options are things that are directly in the simulink code: a,b,c etc. How can I optimize the output of my model via changing Q and f?

Thanks!

[Sai]

Hi Jasmin,

Can you please send me the model and code so that I could also try from my end?

[Jasmin Falconer]

Apoliges for the very delayed response. What is the best way for me to send you the model? I have attached a simplified version of the model and the associated matlab script. I would like to be able to optimize the matlab file parameters in order to maximize the simulink model outputs and then read the output from the simulink.

Answer 1

[Sam Chak]

Hi @Jasmin Falconer

Put it another way around, you can fminunc() to directly optimize the gains from the Simulink model and then probably solve for Q and f from the optimal gains using fsolve().

Else, if the MATLAB script is simple enough that it consists of some straightforward static equations like

a = fun1(Q, f)
b = fun2(Q, f)
c = fun3(Q, f)

then you can describe that in the Simulink model using the fundamental blocks, and use fminunc() to optimize Q and f.