Parameter Identification using Least Square Method Optimization

Hi,

I trying to identify the parameters (di) of a function D(t) = d1*exp(-t) + d2*exp(-t)..di*exp(t) from discrete values D(t_n).

Thus, for n values of t, I have n experimental values of D(t_n).

I want to find the (di) parameters of D(t) by minimizing

F = D(t)-D(t_n), along the n values of t.

F will then be a vector where each component have i unknowns.

I have tried using lsqnonlin function but it doesn't work...

Since D(t) is symbolic, F is symbolic too.

Thanks!

[Andrew Newell]

Are you sure you have written your function right? The first two terms add up to (d1+d2)*exp(-t), which is effectively one parameter.