Abstract
In this paper we present an approach to the solution of parameter estimation problems in systems described mathematically by differential algebraic equations. The numerical solution of these equations is compared with measurements from experiments. By adaptation of the parameters in the differential algebraic equations we try to fit the solution to the measurements. For the fit we use a (weighted) least squares criterion. Not only the final estimate of the unknown parameters, but also additional information about the reliability is derived. A computer program, with an interactive graphical user interface, has been developed to steer through the computation in order to influence the precise formulation of the model and the data and to see the numerical results by direct visualisation.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have