optimization of location using GA

hello guys i really need help I'm trying to distribute circles within location area so the distribution offer less overlap and less holes between the circles my fitness function codes are

function [fit] = cost_xy_ary_by_Dist(xy_cor,r,area,w1,w2)
w1=0.3;
w2=0.7;
[row col]=size(xy_cor);
cost=0;

for i=1:row
    cost(i)=0;
    out(i)=0;
    x1=xy_cor(i,1);
    y1=xy_cor(i,2);
   % xy_cor(i,:)=[]

      a=  ( sqrt(  (((xy_cor(:,1) - x1).^2 ))  +  (((xy_cor(:,2) - y1).^2 )) ));
      if(a<2*r)
          cost(i)=cost(i)+1;
  %     cost(i)=cost(i)+( sqrt(  (((xy_cor(:,1)-1) - x1).^2 )  +  (((xy_cor(xy_cor(:,2) - y1).^2 )) )- 2*r));

%     ov= find( ( sqrt(  (((xy_cor(:,1) - x1).^2 ))  +  (((xy_cor(:,2) - y1).^2 )) ))<2*r);
%     if(find(ov==i))
%        ov(find(ov==i))=[];
%     end
    else
    if(a>2*r)
        out(i)=outter(i)+1;
    end
    end    
    %pdecirc(x1,y1,r)     
end
cost_all=cost./area;
outter=out./area;
fit=w1*cost_all+w2*outter;
end
but it's doesn't work any one can tell me why i'll be very thankful

Answer 1

[Alan Weiss]

It would help me to understand what you are doing, and might help you, if you would do the following:

• Define exactly what your control variables are, meaning the variables you want to move in order to optimize something. Are there any constraints, such as bounds, on your variables?
• Define exactly, in words or a formula (not code) what the objective function is. I mean, give a formula or exact description so that we can understand exactly what you are trying to minimize.
• After you have given us those two things, show us the code that is supposed to calculate your objective function. Not minimize your objective function, just evaluate the thing as a function of the control variables.
• After all that, maybe you will have the problem well-formulated enough that it will be possible to optimize.

Good luck,

Alan Weiss

MATLAB mathematical toolbox documentation