scientific article; zbMATH DE number 1285208

General biology and biomathematics (92B05) Biochemistry, molecular biology (92C40) Adaptive control/observation systems (93C40) Application models in control theory (93C95)

Review text

This work is organized as follows: In the first chapter we introduce the commonly used mathematical model for biotechnological processes. We construct restrictions, that are sufficient to obtain the realistic properties nonnegativity and boundedness of the state variables. In that context we create the terms ``cyclic and ``non-cyclic process and propose an algorithm that is able to decide, only based on the reaction scheme or the stoichiometric matrix of the process, whether a process is cyclic or not. The assumption of non-cyclicity replaces the common assumption of Conservation of Mass and is sufficient to guarantee that the state variables are bounded. The second chapter deals with the application of a discrete version of the \(\lambda\)-tracker to a system of difference equations, that we obtain by discretizing a continuous nonlinear system with Euler's One-Step method. We prove that the adaptive \(\lambda\)-tracker is able to force the output asymptotically towards the \(\lambda\)-neighbourhood of the given reference signal. Furthermore we analyze a special controller with a slightly modificated feedback law compared to the feedback law normally used for the \(\lambda\)-tracker. We show global stability for the scalar linear case. For the two-dimensional case already difficulties arise, but under certain restrictions we are able to demonstrate boundedness of the trajectories and even local stability. In the third chapter we treat the control of some biotechnological processes under input constraints. First we elaborate restrictive assumptions for the process in order to guarantee a successful control of an external substrate using either the dilution rate or the feed rate of one component as input function. Using methods from the theory of ordinary nonlinear differential equations, we prove that the \(\lambda\)-tracker with input bounds applied on the mathematical model of a biotechnological process is able to control the concentration of an external substrate of the process. We proceed in the same way for the control of an autocatalyst or a product, and we show in examples and counter-examples that the assumptions imposed to the process are realistic and make sense. We illustrate also the discrete-time case with an example and confirm in this way that the discrete-time version of the \(\lambda\)-tracker is absolutely useful in practice.

Recommendations

Cited in

(12)

This page was built for publication:

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4242079)