How to direct the variable in the genetic algorithm function to workspace function?

variable1 = input('Variable 1: ');
variable2 = input('Variable 2: ');
variable3 = input('Variable 3: ');
variable4 = input('Variable 4: ');
function y = myfunc(x)
y = x(1)*variable1 + x(2)*variable 2
function [c,ceq] = constraint(x)
c = [x(1)+x(2)-variable3;x(1)*x(2)-variable4];
ceq = [];
ObjectiveFunction = @myfunc;
nvars = 2;    % Number of variables
LB = [0 0];   % Lower bound
UB = [100 100];  % Upper bound
ConstraintFunction = @constraint;
[x,fval] = ga(ObjectiveFunction,nvars,[],[],[],[],LB,UB,ConstraintFunction);

As shown in the code above, I am trying to optimize the objective function based on some inputs from user. However, the function handler does not permit the variable from workspace direct to the function even their name is same. Is there any solution for this case?

Answer 1

[Stephen23]

You need to parameterize the objective function, which can be achieved either using an anonymous function or nested functions:

https://www.mathworks.com/help/matlab/math/parameterizing-functions.html

https://www.mathworks.com/help/optim/ug/passing-extra-parameters.html

Method one: nested functions: you could do something like this:

function [x,fval] = myga(v1,v2,v3,v4,LB,UB)
[x,fval] = ga(@myfunc,2,[],[],[],[],LB,UB,@constraint);
  function y = myfunc(x)
    y = x(1)*v1 + x(2)*v2;
  end
  function [c,ceq] = constraint(x)
    c = [x(1)+x(2)-v3;x(1)*x(2)-v4];
    ceq = [];
  end
end

and call it like this:

var1 = str2double(input('Variable 1: ','s'));
var2 = str2double(input('Variable 2: ','s'));
var3 = str2double(input('Variable 3: ','s'));
var4 = str2double(input('Variable 4: ','s'));
[x,fval] = myga(var1,var2,var3,var4,[0,0],[100,100])

Method two: anonymous functions: you could do this (with version R2016b or later, where functions may be defined at the end of a script):

var1 = str2double(input('Variable 1: ','s'));
var2 = str2double(input('Variable 2: ','s'));
var3 = str2double(input('Variable 3: ','s'));
var4 = str2double(input('Variable 4: ','s'));
LB = [0,0];
UB = [100,100];
objfun = @(x)myfunc(x,var1,var2);
confun = @(x)constraint(x,var3,var4);
[x,fval] = ga(objfun,2,[],[],[],[],LB,UB,confun);
function y = myfunc(x,v1,v2)
y = x(1)*v1 + x(2)*v2;
end
function [c,ceq] = constraint(x,v3,v4)
c = [x(1)+x(2)-v3;x(1)*x(2)-v4];
ceq = [];
end

[Walter Roberson]

You need to do this for the nonlinear constraint function too.

[Soon Kok Yew]

Works well! Thanks alot!