# How to fit data with numerical solution of system of parametric ODE? [duplicate]

I need to find the parameters (k1,k2,k3,k4,k1r,k2r,k3r,k4r) that fit my data (list of [Intensity, time]) using the function b(t), solution of the following system of parametric ODE:

{a'[t] == -k1*a[t] - k2*a[t] + k1r*g[t] + k2r*p[t],
g'[t] == -k1r*g[t] - k3*g[t] + k1*a[t] + k3r*b[t],
p'[t] == -k4*p[t] - k2r*p[t] + k4r*c[t] + k2*a[t],
b'[t] == -k3r*b[t] + k3*g[t],
c'[t] == k4*p[t] - k4r*c[t]}


Initial conditions:

a[0] == 1,
c[0] == 0,
b[0] == 0,
p[0] == 0,
g[0] == 0,


The system of ODE has no analytical solution, so I tried to use NDSolve, but it is of course not working because the ODEs are parametric. I guess ParametricNDSolve would help me, but unfortunately it is not implemented in the version of Mathematica I am using (Mathematica 8). Do you have any idea about how to address this problem, avoiding the ParametricNDSolve function?

• See Help > Wolfram Documentation > FindFit > Applications > Differential Equations for an example. I will give several MSE links in a followup comment. Commented Jul 31, 2014 at 16:35
• Here are variants of this question, many with responses. I found them using Google and wsearching for (without the quotes) "mathematica stackexchange fit differential". 1 2 3 4 Commented Jul 31, 2014 at 16:37
• More MSE links: 5 6 7 Commented Jul 31, 2014 at 16:38
• As far I see, all these examples use ParametricNDSolve, am I right? I would need an alternative method, because I don´t have Mathematica 9. Thanks Commented Jul 31, 2014 at 20:56
• Several of the links I gave provide responses that use NDSolve rather than ParametricNDSolve. Commented Jul 31, 2014 at 22:57