ABSTRACT
Many pharmacokinetic models produce nonlinear equations. Drug concentrations versus time curves after intravenous (IV) bolus, IV infusion, or oral administration are all nonlinear. Using nonlinear regression analysis it is possible to minimize the Weighted sum of squared across all the data points in a somewhat more objective fashion. The grid search method could never be considered an efficient nonlinear regression method, however, it does illustrate the general principles involved in optimizing parameter values. D. W. Marquardt modification attempts to use the best of the steepest descent method and the Gauss-Newton approach. An approach that has become quite popular in nonlinear regression analysis is the simplex method. The Damping Gauss-Newton method uses the same equation as the ordinary Gauss-Newton method. Common to each of the nonlinear regression methods is the potential to fail by converging to a local minimum away from the best or global minimum.
