Optimization of biotechnological processes. The acetic acid fermentation. Part II: Practical identifiability analysis and parameter estimation

Abstract

In part I of this series a mathematical model for acetic acid fermentation was reported. However, no kinetic model can be complete until its equation parameters are estimated. This inevitably entails a practical identifiability analysis intended to ascertain whether the parameters can be estimated in an unambiguous manner based not only on the sensitivity of the model to them, but also on the amount and quality of available experimental data for this purpose. Also, estimating the model parameters entails optimizing a specific objective function subject to the model equations as major constraints and to additional, minor constraints on variables and parameters. This approach usually leads to the formulation of a non-linear programming problem involving differential and algebraic constraints where the decision variables constitute the parameter set to be estimated. In the scope of modelling biotechnological processes, this problem is not usually dealt with in a proper way. This second paper reviews available models for practical identifiability assessment and parameter estimation with a view to their prospective application to the proposed model and its validation.