How to save each result of objective function solved by genetic algorithm for each iteration?

Hello!

• I have solved my optimization problem using genetic algorithm.My objective function is related to a random matrix that's why it is changed in each execution of my program.I want to put a specific number of iterations and store the result of my objective function of each one.It is possible to do that?
• any suggestion is appreciated

Answer 1

[Walter Roberson]

Yes, of course.

numiter = 50;
results = zeros(nvars, numiter);
fvals = zeros(1, numiter);

for iter = 1  : numiter
    [results(:, iter), fvals(iter)] = ga(...);
end

[bestfval, bestidx] = min(fvals);
bestx = results(:,bestidx).';

[Maria]

@Walter Roberson sorry for my late reply, but "ga" takes a lotof time in order to solve my problem and at the end it doesn't give the required result.

This is my code , could you please show me where should the loop statement

function [sol,fval,exitflag] = MyProblem()


% all my variables are here

%=====================================
 
 data1 = load('all_N1.mat');   % i upload the random matrix, 
 all_N1 = data1.all_N1;
 data2 = load('all_qm1.mat');  % qm1 is related to N1
 all_qm1 = data2.all_qm1;
 
 Num = 5;     % the number of iteration 
 
sol = zeros(1,Num);
fval = zeros(1,Num);

 for iter = 1 : Num
     
 qm = all_qm5{iter};
  N = all_N5{iter};

 a = floor(L/(UAV_Speed*dt));
 q_int = optimvar("q_int","Type","integer", "LowerBound",0,"UpperBound",a);
 S = optimvar("S",[M,1],"Type","integer","LowerBound",0,"UpperBound",1);

obj = fcn2optimexpr(  @Objective_function   ,S);
 
 prob = optimproblem('ObjectiveSense','max');
 prob.Objective = obj ;

 constr1 = fcn2optimexpr(@NLCON_MAIN,S,q_int)<=0;
 constr2 =  cumsum(S.*(sum(Energy_Pr,2))) <= E_max(1)-E_prop;

 prob.Constraints.constr1 = constr1;
 prob.Constraints.constr2 = constr2;


 [sol(:,iter),fval(:,iter),] = solve(prob) ; %my problem is solved with "ga"

 
 end 
 [bestfval, bestidx] =  min(fval);
 bestx = sol(:,bestidx).';
 

Also i don't know if there is a solution to accelerate "ga" execution.

[Walter Roberson]

Your code is already running Num iterations, each with a different qm and N value . However, you do not use qm or N so all of the iterations are running the same problem.