Robust model predictive online control of bioproduction processes based on fungal growth inspired multi-objective optimisation algorithms

Since millennia, microorganisms have been used in the production of

foods and beverages. More recently, microorganisms have also

shown their potential as producing agents of fine chemical

compounds. However, while most microorganisms are incapable of

producing (an economically viable concentration of) a targeted

compound, they are often genetically modified for this purpose. The

studied process in this project is the production of the sweet protein

thaumatin-II with the use of a genetically modified Pichia pastoris

GS115 strain. A first goal of this project is the development of a multi-scale kinetic

metabolic model of the used P. pastoris strain. A major emphasis will

be laid on understanding the regulatory mechanism of the inserted

thaumatin-II sequence during the different process phases. This will

result in a better understanding of the production process and how to

regulate it. The eventual goal of this project is the use of the

developed metabolic network for the online optimal control of the

thaumatin production process. The use of a novel hybrid optimisation

algorithm, based on mushroom-growth, will allow for the agile

generation of well-converged optimal working points based on online

measurements and corresponding model-based predictions.

The anticipated results of the research will intensify metabolic

network integration in biotechnological process optimisation, with an

emphasis on shifting from open loop control mechanisms towards

online optimal control systems.