How can i use Genetic algorithm in simulink for PID?

Hello,

I want to use genetic algorithm to find Kp,Ki and Kd for PID controller in SimulinK,

can someone give me some ideas ?

Answer 1

[Ibrahim Seleem]

The main idea is creating s fitness function which depend on the process you want. In your case, I suggest to use sum of square error

function y=fcn(x)

Kp=x(1);

Ki=x(2);

Kd=x(3);

simopt = simset('solver','ode5','SrcWorkspace','Current','DstWorkspace','Current'); % Initialize sim options

[yout]=sim('your model name',[0 simulation time],simopt);

err1=error.^2;

err=sum(err1);

end

This is a simple code, if you can upload your complete model, I can help you to modify it.

[Sam Chak]

@AHMET SUNGUR ALP SARGIN

Can you open and ask a new question?

fitfun  = @hata;
options = optimoptions('ga', 'ConstraintTolerance', 1e-6, 'PlotFcn', @gaplotbestf);
[K, fval] = ga(fitfun, nvars, A, b, Aeq, beq, lb, ub, nonlcon, options)

K =

1.37333774469592 1.13832193717631 0.392244213118769

fval =

1.67188417949144

Answer 2

[JESUS DAVID ARIZA ROYETH]

To perform a genetic algorithm in simulink and find optimal parameters of a pid controller, the first thing you have to do is define how you are going to encode these parameters in a population, I recommend that the population encode it in digits, after this, you must define the performance of each individual based on the stability of the process (which you can evaluate by simulating some disturbances) and thus being able to select individuals in several iterations, you will probably need help from some codes in Matlab, since it would be much easier to do it in Matlab directly