how can i write cost function in firefly algorithm for optimal tuning pid parameters in matlab

Iam using firefly algorithm for optimal tuning pid parameters , the cost function is IAE i wrote IAE function as a function of pid parameters {kp,ki.kd} but it doesnot work it gives error that it undefined variable ,should i give initial values for parameters?? and what is error in the IAE function

   costFunction = integral(abs(error),0,Inf);

it gives Undefined function 'abs'

Answer 1

[noor diyana]

Here are example from my senior project. You can try it. You can change the objective function, MaxIt, runtime as system that you want. Anything can email me [e-mail] All the best!

[sandy jeak]

Thanks for answering me, I put my system in the code and i have that error Reference to non-existent field 'Position'. Error in main (line 112) Trace_pid=[Trace_pid;BestSol.Position]; How can I fix that. Thanks in advance

Answer 2

[ednilson szeskoski]

I have the same problem. I'm implementing PID for a voltage control system. The goal is to optimize using firefly, the same error quoted is happening, how can I implement in the code the transfer function of my case? Thank you.

tf([0.004, 0.454, 5.55, 15.1, 10], [1.6e-05, 0.002032, 0.04732, 0.4286, 2.133, 8.76, 18.01, 11])

[Walter Roberson]

I am confused about whether you are using Simulink or trying to do it in MATLAB itself? Are you trying to implement Firefly algorithm in Simulink? Are you trying to implement Firefly algorithm where the function being computed is the result of simulating a Simulink model?

[ednilson szeskoski]

The image where the mesh is drawn representing the system is only for understanding my system I work with it in simulink but only by changing the PID values, the image with the equation is the result of the transfer function of the feedback system, in which I am not able to implement in the Diyana code.

Below I have the code that I made simply to get the transfer function. Now the problem is to apply the adapted PID to the firefly optimization code.

%%Malha da excitação

amp = tf([10] , [0.1 1])

exit = tf([1] , [0.4 1])

ger = tf([1] , [1 1])

sens = tf([1], [0.01 1])

h = amp * exit * ger

g = sens

malhaeq = h / (1+h*g)

step(malhaeq)

[ednilson szeskoski]

https://www.mathworks.com/matlabcentral/answers/334697-how-can-i-implement-a-pid-controller-with-firefly-algorithm-to-optimize-pid-parameters