Control and modelling of bioprocesses

The Bioprocess Control and Modelling research group is interested in the advanced control, monitoring and modelling of processes that use populations of living microorganisms to produce useful substances such as biopolymers or antibiotics. Methods for information mining, methods for extracting knowledge from different types of data and information, and methods for building models of bioprocesses, including physiological ones, are investigated. In the area of control, intelligent methods for physiological control of bioprocesses such as knowledge-based control, fuzzy online classification of physiological states or artificial neural networks as software sensors are studied. Specific model bioprocesses include e.g. production of mcl-PHA biopolymer by Pseudomonas putida strain or production of antibiotic nystatin by Streptomyces noursei strain.

Areas of research

• application of information extraction methods and methods for extracting knowledge from different types of measurements used in bioprocesses (time courses, frequency spectra)
• application of knowledge obtained by information data mining methods and current knowledge of metabolic engineering to the creation of physiological models of bioprocesses
• study of intelligent methods for the physiological control of bioprocesses such as knowledge-based control, fuzzy on-line classification of physiological states or artificial neural networks for models and software sensors
• hybrid and adaptive software sensors for advanced monitoring of bioprocesses
• development of a comprehensive software simulator for fed-batch yeast cultivation
• use of reinforcement learning methods in bioprocess modelling and control