Reinforcement Learning for Process Optimisation and Control

Sustainable production of bioenergy by microalgae has received considerable attention since the end of the 20th century as it can directly utilise solar energy and CO2 to synthesise biorenewables. However, due to the lack of process efficiency, bioprocesses are still economically inviable. Reinforcement Learning (RL) is a subfield of Artificial Intelligence (AI) which trains "machines" to make optimal decisions. RL has made great achievements, such as "machines" self-learning to play board games or videogames. We will make use of this rapidly-growing technology to address current challenges faced by bioprocesses, by adopting RL to optimally design, control and operate bioprocess production.

The proposed research is highly significant and rather topical in the wake of the Covid-2019 pandemic. Indeed research into RNA vaccines can reduce the time to respond to disease outbreaks from years to weeks. This PhD project will develop the tools and techniques to overcome major hurdles to the implementation of physical models in the process engineering of vaccines, such as the lack of bioanalytical tools for accurate measurements.