OONumerics User :

From: Arne Dieckmann (mail_at_[hidden])
Date: 2008-02-19 04:06:40


Hi all,

I'm pretty new to the field of numerics. I'm developing a Java application which should be able to solve a system of differential equations and fit certain parameters to experimental data. For example, the ODE could look like f'(a) = k*a*b, where a and b are variables for which I have the experimental data and k is the parameter which should be fit. I know how to solve differential equations (Euler, Runge-Kutta,...) numerically and how to do a polynomial approximation. But this seems somehow to be a combination of both things, right? I would have to integrate the ODE first and then do the approximation, but I don't know how to do the numerical integration if I have an unknown parameter k. I wondering if I'm just confused by the problem or if I need a new strategy. Thanks in advance!

Fret