Fmincon with multiple linear constraint of multidimensional array

Hi,

I'm solving a minimization problem that requires a variable that is a function of parameters to be positive. However, this variable is in fact a matrix and I would like to impose the constraint using the fmincon alternatives. The problem is as follows: the matrix of interest is g=repmat(X*b(1:10,1),1,10)+b(11)*d; where X is data, b is the parameter vector and d is a matrix of size(N,10) where N is the number of observations. d is a matrix of (10) random draws of a standard normal for each observation. Therefore, g is a matrix of size N times 10. I need to guarantee that the parameters give only values of g greater than zero and at the same time I don't want to include something such as g=max(repmat(X*b(1:10,1),1,10)+b(11)*d,eps); because that could make fmincon to search for values of b that don't make sense.

If I only had one value for d, this would be the standard problem of a linear constraint and I would just need to define the A and b components as explained in the documentation. However here I have 10 of these and I don't know how to pass that to fmincon.

Any ideas?