LINPROG requires the following inputs to be of data type double: 'f'.

%% Discrete modelling 
Ae=(1-(K*ts)/Cth);
B=(ts*COP*HP)/Cth;
D=ts*K/Cth;
u=zeros(1,i);
Tin=zeros(1,i);
x=zeros(3,i);
cost=[gamma;zeros(1,1008);zeros(1,1008)];
xcost=zeros(3,i);

for j=1:i
    
     if j>1
     Tin(j)=Ae*Tin(j-1)+B*u(j-1)+C*Tout(j-1);
     x=[u(j-1);Tin(j-1);gamma(j-1)];
    else
        Tin(j)=0;
        u(j)=0;
        x(j)=0;
     end
    
%     x=[u(j);Tin(j);gamma(j)];
    cost=[gamma(j);0;0];
    f=@(x) sum(cost(j).'*x(j));
 
    %Starting point%
    x0=0;
    %Constraints%
    %Inequality const%
    A=[1 0 0;-1 0 0;0 1 0;0 -1 0;0 0 1;0 0 -1];
    b=[umin; -umax; Tmin; -Tmax; gammamin; -gammamax];
   
    
    options = optimoptions(@linprog,'Display', 'off');
    
    [x(j),xcost(j)]=linprog(f,x0,A,b);

end    

[Ricardo López]

Ae, B and D are constants

Answer 1

[Ameer Hamza]

linprog() does not require you to multiply 'f' with x. It will do that internally. Just directly give the vector 'f'

f = cost; % instead of f = @(x) sum(cost(j).'*x(j));

[Ricardo López]

I am still getting this error: The number of rows in A must be the same as the number of elements of b.

And I also tried to solve this by fmincon with f = @(x) sum(cost(j).'*x(j));

[Ameer Hamza]

The dimensions of A and b seem to be fine? Can you add a breakpoint and see if the dimensions are correct when the issue happens?