|Frank Solina Gilbert|
Hi yehuda :),
Thanks for replying to my request.
Your are right I really should use A,B,C,DD as variables, but the code really works for the first case. Because I have the data set, and yes normally you have to define A,B,C,DD before you apply NDSolve . But I have the my idea from the very same wolfram forum. Here is the URL: http://support.wolfram.com/mathematica/mathematics/stats/ndsolvefit.html.
Believe me, it really works, because this type of code enables Mathematica to start with own initial guesses. You even don't need to give the variables A,B,C,DD before. The problem I have is the second kind of the code. Because I have only one differential equation and 2 unknowns u[t] and g[t]. That would mean that I have to define g[t] brfore. But how??? g[t] is the input to the system whereas the u[t] describes the system's response. g[t] is a randomly generated function, defined only by the experimental conditions I have. Is there a way that I at least can define or describe a little bit g[t], before using it as an input.