Trouble creating function handle with matrices in matlab

I have a matrix x1 and I am trying to use nlinfit to eventually obtain values of b, b(1) and b(2).

I believe that my function handle is incorrectly doing matrix multiplication, however I cannot figure out what I must do to fix it.

x1 = 0:0.05:1;
x2 = 1-x1;
T = 313.15;
R_const = 8.314;
mdlfunc = @(b,x1) -R_const*T(x1.* log(x1 + b(1).* x2) + x2.* log(x2 + b(2).* x1));

A = 1286.46130;
for i = 1:numel(x1)
    gE_exp(i) = A*x1(i)*x2(i);
    
end


b0 = rand(1,2);
[b,R,J,CovB,MSE] = nlinfit(x1,gE_exp,mdlfunc,b0);

my error message:

Error using nlinfit (line 213)
Error evaluating model function '@(b,x1)-R_const*T(x1.*log(x1+b(1).*x2)+x2.*log(x2+b(2).*x1))'.

Error in Proj1B (line 28)
[b,R,J,CovB,MSE] = nlinfit(x1,gE_exp,mdlfunc,b0)

Caused by:
    Array indices must be positive integers or logical values.
 

Answer 1

[Steven Lord]

mdlfunc = @(b,x1) -R_const*T(x1.* log(x1 + b(1).* x2) + x2.* log(x2 + b(2).* x1));

You meant to multiply T by the quantity (x1.* log(x1 + b(1).* x2) + x2.* log(x2 + b(2).* x1)) but the way you wrote it you indexed into the (scalar) T with that quantity, which is unlikely to always be equal to 1.

[Conor Luppnow]

wow, I am kicking my self right now. Thank you very much.