A genetic algorithm with decimal coding for the estimation of kinetic and energetic parameters

A hybrid estimation method is proposed to estimate simultaneously the kinet ic and energetic parameters of chemical models. This method combines a gene tic algorithm with a local convergence method, the former generating initia lisation points for the latter. The method is applied to a real complex che mical system: the saponification of ethyl acetate at a concentration of 1 m ol 1(-1). The reaction is carried out in a calorimetric reactor which deter mines experimental profiles of power delivered or absorbed by the reaction mass. A model, based on mass and energy balances in the reaction mass, incl uding the kinetic and energetic parameters of the reaction, allows to calcu late: the power profile. The comparison between the experimental and calcul ated profiles gives the criterion to minimise. The parameters of the model are successfully determined. The hybrid method determines more efficiently and more rapidly the solution of the problem than a set of several estimati ons with a local convergence method. (C) 2000 Elsevier Science B.V. All rig hts reserved.