Using fmincon with for loop

Hi!

I have a minimalization problem, where i have 3 variables (x1, x2 and x3) and two non linear equations:

x(1)+parama*x(2)*x(3)=Ptloademod(k)

paramb*x(2)+(paramc-paramb)*x(2)*x(3)=Ptloadh(k)

Where parama, paramb and paramc are parameters defined earlier in the code. x1 x2 are >=0 and 0<=x3<=1. Ptloademod(k) is the k-th element in a predefined row vector, the same can be said about Ptloadh(k). The two row vectors contain 96-96 elements.

My objective function is:

min c1*x1+c2*x2 (where c1 and c2 are predefined constants).

I know that i have to use nonclon:

function [c,ceq]=myf(x, Ptloademod, Ptloadh,k,paramb,paramc,parama)

c=[];

ceq(1)=x(1)+parama*x(2)*x(3)-Ptloademod(k);

ceq(2)=paramb*x(2)+(paramc-paramb)*x(2)*x(3)-Ptloadh(k);

end

And after that I have to use fmincon:

x0=[1,1,1];

A=[];

b=[];

Aeq=[];

beq=[];

lb=[0,0,0];

ub=[Inf,Inf,1];

nonlcon =@myf;

x=fmincon(objective,x0,A,b,Aeq,beq,lb,ub,nonclon);

So I would like to use a for loop to find the optimal solution for all the 96 cases and create 3 new row vectors containing the x1 x2 and x3 for the 96 cases.

Hope you can help me :)

Answer 1

[Alan Weiss]

I think that you need to define nonlcon this way:

nonlcon = @(x)myf(x, Ptloademod, Ptloadh,paramb,paramc,parama);

Notice that I removed your k argument.

And you need to put that definition inside your loop. Something like this:

N = length(Ptloademod);
solutions = zeros(N,3);
for i = 1:N
    nonlcon = @(x)myf(x, Ptloademod(i), Ptloadh(i),paramb,paramc,parama);
    x = fmincon(objective,x0,A,b,Aeq,beq,lb,ub,nonclon);
    solutions(i,:) = x;
end

Alan Weiss

MATLAB mathematical toolbox documentation