Fitness Function In the PSO algorithm

Question

Maria 2014-5-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/129742-fitness-function-in-the-pso-algorithm

回答： Walter Roberson 2018-12-31

My problem it is about PSO with PID controller , i just tried to tune PID parameters with PSO for tracking circular path as input (input is circular path ) with desired velocity = 0.25 [meter/sec] and desired azimuth given by = (2 * 3.14 * f(t)/m) rad , with m = -40 where m is the slop, f(t)=t , 0=t=40 and sampling time = 0.01 sec . I wrote the code of PSO but it does not work , I think the problem is in the fitness function structure of PSO algorithm and I don't know what is the bugs in the code below ,the fitness function should compute the Mean Square Error , please I hope from you to help me .

In the picture below the desired input is a circular path with a above parameters and the actual output is should be as the desired input .

Many thanks

the fitness function that I wrote :

function F = FitnessFun(pid)

         Kp1 = pid(1);
         Ki1 = pid(2);
         Kd1 = pid(3);
         Kp2 = pid(4);
         Ki2 = pid(5);
         Kd2 = pid(6);
         sprintf('The value of interation Kp= %3.0f, Kd= %3.0f', pid(1),pid(2),pid(3),pid(4),pid(5),pid(6)); 
         % Compute function value
         simopt = simset('solver','ode5','SrcWorkspace','Current','DstWorkspace','Current');  % Initialize sim options
         [tout,xout,yout] = sim('sim1',[0 100],simopt);
         n=1;m=-40;
for t=0.3:.3:2*pi
    theta_d(n)=((2*pi)* t/m);
    n=n+1;
end
t=1:20;
v=step(1,t);
[theta,t]=lsim(theta_d,t);
for k=1:20
    eth(k)=theta(k)-theta_d(k);
end
ev=(-v+1);
out=sum((ev.^2)+sum(eth.^2))/20;
        F=out; 
    end