problam with genetic algorithm individuals intialization

I'm working with genetic algorithm to optimize umts network at first i have BTS distribute randomly each has (x,y) coordinates i generate the candidates solutions using rand the solutions are numbers to be summed with the original x and y they i have tried this

clc
    clearvars
    ver_code=1;             % Version of code -> for reference only
    ga=1;                   % Counter
    max=2000  ;               %@ Maximum number of generationss
    radius=2.5;            %@ The exact value is 0.35
    N=31;                   %@ Number of BS`s
    limits=30;               %@ The limits of the coverage area in Kilometer
    pop_size=300;          %@ The population size
    % co_prob=0.8;
    mutrate=0.01;            %@ The Mutation rate
    selection=0.5;          % The persentage of survival 
    Generation_index=0;     %-> for reference only
    %__________________________________________________________________________
    load('C:\Users\zaizafoon\Desktop\three.mat','ua');    % create random points
    [xa,ya]=size(ua);
    pop_x=((rand(pop_size,N)*limits)); % create random populatio for X points
    pop_y=(((rand(pop_size,N)*limits))); % create random populatio for Y points

      all_cost_ary=zeros(1,pop_size);     % Initialise an array specifing the cost for each population -> for reference only
      prev_cost=-1000;                     % This variable is used only for finding the best population;

      %M=ceil((pop_size-keep)/2);         % Number of parents
      keep=floor(selection*pop_size);     % number of mating
      odds=1;
      for ii=2:keep
      odds=[odds ii*ones(1,ii)];
      end
      Nodds=length(odds);
      odds=keep-odds+1; 

      %__________________________________________________________________________

      while ga<max

          ga=ga+1

          cost=zeros(1,2);
      %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
      %cost calculations

      for t=1:pop_size
             for i=1:N
                 %Shifting 
              test_ary(i,1)=ua(i,1)+pop_x(t,i);
              test_ary(i,2)=ua(i,2)+pop_y(t,i);
              if(test_ary(i,1)>limits)
                  test_ary(i,1)=test_ary(i,1)-limits;
              end
              if(test_ary(i,2)>limits)
                  test_ary(i,2)=test_ary(i,2)-limits;
              end
             end

        cost_per_pop(t)=cost_xy_ary_by_Dist(test_ary,radius);
        if prev_cost<cost_per_pop(t)
            prev_cost=cost_per_pop(t);
            best_cost_pop_x=pop_x(t,:);
            best_cost_pop_y=pop_y(t,:);
            Generation_index=ga;
          end
but the candidates solution were all the same so the fitness function does not change for each individuals what I'm doing wrong ?