Problem with index in position 1 exceeding array bounds in PSO algorithm

Hello guys,

I have a particle swarm optimization code in Mtalab that should find the local minima of a function with two variables. As of now the code is not running because I am getting the error: Index in position 1 exceeds array bounds.

Any thoughts on how to fix this? (I know this is a usuall beginner error but I am just that, a beginner)

I also have another question, anyone knows how to do the following in a if loop or such? I want to multily the weight w with a factor close to 1 such that the multpilication occurs with every iteration till the weight hits a lower limit of 0.3 and then it won't go below that.

This is is my script (I have another scrip for the main objective function):

x_min=[-5,-5];
x_max=[5,5];
n=100;
m=2;
c1=2;
c2=2;
w_max=1.4;
w_min=0.3;
B=0.95;
v_max=5;
Max_iter=100;
for i=1:n
    for j=1:m
        pos(i,j)=x_min(i,j)+rand().*(x_max(i,j) - x_min(i,j));
    end
end
vel=pos;
%function evaluation
out=fun(pos);
pos_best_val=out;
pos_best=pos;
[fminval,index]=min(out);
glo_best=pos_best(index,:);

X=pos;
Out=fun(X);

k=1;
w=w_max;
while k <= Maxiter
    w(k)=w_max*B;
    if w<w_min
        w=w_min;
    end
    %Update values and positions
    new=find(out<=pos_best_val);
    pos_best(new,:)=X(new,:);
    pos_best_val=out(new);
    
    [f_best_val,ind1]=min(pos_best_val);
    if f_best_val<=f_min_val
        f_min_value=f_best_val;
        glo_best=pos_best(ind1,:);
    end
    
    for i=1:n
        for j=1:m
            vel(i,j)=w.*vel(i,j)+c1.*rand().*(pos_best(i,j)-pos(i,j)) + c2.*rand().*(glo_best(1,j)-pos(i,j)); %updating velocity
            pos(i,j)=vel(i,j) + pos(i,j); %updating position
            %Handling boundary conditions 
            if pos(i,j)<x_min(j)
                pos(i,j)=x_min(j);
            elseif pos(i,j)>x_max(j)
                pos(i,j)=x_max(j);
                %Velocity condition
                if vel(i,j)> v_max
                    break
                end
            end
        end
    end
    
    k=k+1
end
Minima(run)=fun(glo_best);
Minima_pos(run,:)=glo_best;

Answer 1

[the cyclist]

x_min and x_max are 1x2 vectors, but in the loop you try to access a second row of those matrices, which do not exist.

I'm not sure what you intend to happen there, so I can't give advice.

[Christian Tsobanoglou]

I see, thank you for the answer!