Evolutionary Algorithms for Static and Dynamic Optimization of Fed-batch Fermentation Processes

Abstract

In this work, Evolutionary Algorithms (EAs) are used to control a recombinant bacterial fed-batch fermentation process, that aims to produce a bio-pharmaceutical product. Initially, a novel EA was used to optimize the process, prior to its run, by simultaneously adjusting the feeding trajectory, the duration of the fermentation and the initial conditions of the process. Finally, dynamic optimization is proposed, where the EA is running simultaneously with the fermentation process, receiving information regarding the process, updating its internal model and reaching new solutions that will be used for online control.

Experiments on Computer Assisted Optimization of the Escherichia Coli Fermentation Process Using Optferm

“…We studied the state of the art to evaluate the available tools in order to select the most appropriate one for the purpose of escherichia coli fermentation optimization [5] [6]. The selected tool was Optferm, presented and described in [7].…”

“…The model and the initial parameters used in this work were published in [6] and later incorporated in Optferm. Thus, the state variables used and the initial kinetic parameters are the result of experimental work developed in real conditions and are crucial to the quality of the obtained results.…”

Continuous Dynamic Constrained Optimization—The Challenges

“…The consequence of this dynamic/adaptive scheme is that if the detector solutions used by the change detection method are infeasible or become infeasible, over time their fitness value will decrease. 7 In some other CH techniques which use ranking-based methods [38]- [40], during the selection process the current better solutions might be dropped in favor of other solutions, which might have worse fitness values but are more useful for the CH process. In these situations there might also be a drop in the values of the best solutions at each generation.…”

“…In real-world DCOPs, the objective function and constraint functions can be combined in three different types: 1) both the objective function and the constraints are dynamic [2]- [4]; 2) only the objective function is dynamic while the constraints are static [5]- [7]; and 3) the objective function is static and the constraints are dynamic [8]- [10]. In all three types, the presence of infeasible areas can affect how the global optimum moves, or appears after each change.…”

Modelling of Biotechnological Processes – An Approach Based on Artificial Neural Networks

“…Furthermore, it is one of the best characterized systems regarding molecular genetics, physiology and expression systems. An important problem in this culture lies in the tight control that is necessary during the fermentation process, since at particular conditions (excessive or low glucose concentration) recombinant protein production can be significantly reduced [25], due to the effect of acetic acid or cellular starvation. The main optimization objective in this process is to control the feeding trajectory in a manner that concentration of glucose maintains at optimum levels for cell growth and subsequent high protein production.…”