How can i improve the performance of Genetic algorithm ?

i have a code using genetic algorithm ,&to get the best performance for 3 anntenas the desired signal at theta =30 & the unwanted signal at=-60 ,my fitness function is to get minimum error,the antenna weights are complex the problem is the error which output from the algorithm is very small but the main desired beam not get on theta=30,i tried to increase the number of iterations & the number of population & put initial point but the performance isn`t good the general algorithm : if true clear clc close all global m m=1; thetaS =30; ObjectiveFunction = @newone1; nvars =6; % Number of variables LB = [0 0 0 0 0 0]; % Lower bound UB = [1 1 1 1 1 1]; % Upper bound d = .5; % element spacing in terms of wavelength d = lambda/2 options = gaoptimset('PopulationSize',300,'CreationFcn', {@gacreationuniform},'FitnessScalingFcn', {@fitscalingprop},'SelectionFcn',@selectionroulette,'MutationFcn',@mutationadaptfeasible,'CrossoverFcn',@crossoverheuristic,'FitnessLimit',0.0001,'Generations',100);%'InitialPopulation',X0);

options = gaoptimset(options,'PlotFcns',{@gaplotbestf,@gaplotstopping,@gaplotbestindiv}, ... 'Display','iter');% Next we run the GA solver.%@gaplotrang

[x,fval,exitFlag,Output, population,scores] = ga(ObjectiveFunction,nvars,[],[],[],[],LB,UB,[],options) theta = -pi/2:.01:pi/2; AF = zeros(1,length(theta)); x=[x(:,1)+j*x(:,2) x(:,3)+j*x(:,4) x(:,5)+j*x(:,6)];

% Determine the array factor for linear array for i = 1:nvars/2

   AF = AF +x(i)'.*exp(1j*(i-1)*2*pi*d*sin(theta)); 
end 

end

the fitness function is called( newone1) & i use least minimum sqaure error : if true % code function abse =newone1(w) d = .5; % element spacing in terms of wavelength d = lambda/2 N =6; % number of elements in array global m

thetaS =30; % input(' What is the desired users AOA (in degrees)? '); thetaI =-60; % input(' What is the interferers AOA(in degrees)? '); %----- Desired Signal & Interferer -----% T=1E-3; t=(1:1000)*T/1000; S=cos(2*pi*t/T); thetaS = thetaS*pi/180; % desired user AOA I = randn(1,1000); thetaI = thetaI*pi/180; % interferer AOA

%----- Create Array Factors for each user's signal for linear array -----%

vS = []; vI = []; i=1:N/2; vS=exp(1j*(i-1)*2*pi*d*sin(thetaS)).'; vI=exp(1j*(i-1)*2*pi*d*sin(thetaI)).';

%----- Solve for Weights using LMS -----%

 m=m+1;
 b=floor (m/301)+1;
   x = S(b).*vS + I(b).*vI;
        w;
    w1=[ w(:,1)+j*w(:,2) w(:,3)+j*w(:,4) w(:,5)+j*w(:,6) ];

    y=w1*x;
     e =(S(b))- y;

    abse=abs(e);
end
  end

[Walter Roberson]

http://www.mathworks.com/matlabcentral/answers/13205-tutorial-how-to-format-your-question-with-markup