Bruno Morabito successfully defended his PhD thesis!
Research focused on “Risk-aware and Robust Approaches for Machine Learning-supported Model Predictive Control for Iterative Processes”
The IMPRS program welcomed Bruno Morabito in 2016, and he successfully defended his PhD thesis on December 20, 2023. Considering the entire journey, he found the IMPRS program to be particularly enriching, pointing out its appeal due to the numerous opportunities offered, such as engaging classes, establishing connections with fellow Ph.D. scholars, and presenting findings at various gatherings and conferences hosted by the program.
As the IMPRS ProEng is a collaborative initiative between the Max Planck Institute for Dynamics of Complex Technical Systems and the Otto von Guericke University (OvGU) in Magdeburg, Bruno Morabito expressed his gratitude towards the OvGU for providing him with the opportunity to attend conferences, which played a pivotal role in broadening his academic horizons.
Under the guidance of Prof. Rolf Findeisen, the research focused on model-based optimal control strategies that account for model uncertainties. Specializing in the application of these methods to bioreactor systems, Bruno Morabito's work aimed to enhance model predictive control (MPC) in iterative processes through the integration of machine learning techniques. It proposes two methods for data-driven uncertainty modelling in iterative processes: one using Gaussian processes for learning model uncertainty and neural networks for the nominal model, and the other based on tube-based model predictive control for ensuring constraint satisfaction. The thesis also introduces HILO-MPC, a novel Python library for implementing machine learning-supported MPC, which interfaces with TensorFlow and PyTorch and offers tools for control and estimation problems using machine learning models. The work aims to improve process performance and safety by incorporating machine learning into predictive control strategies.
The IMPRS program, with its professionalism and organized structure, played a crucial role in preparing Bruno Morabito for his future career. He believes that the IMPRS program excels in providing assistance and direction.
Collaborations during the PhD journey were primarily within the research group and extended to industry, including notable interactions with companies such as Baker Hughes. Despite facing challenges, Bruno Morabito did not seek furhter support from the IMPRS but acknowledged the positive impact of collaborations on their research.
The program offered unique opportunities, including an unexpected interview offer from MathWorks during a conference. Summer schools and conferences emerged as valuable motivators, fostering a sense of community among PhD students.
For future PhD students, Bruno Morabito has the following advice: “If you want to finish quickly, read a very good thesis that was written in your field. Decompose it to understand which components make the thesis and work towards these components. Do not let anything else distract you. But, if you are not in a rush, enjoy exploring different directions.”
In conclusion, Bruno Morabito extends appreciation for the program's support and leaves a final note of thanks for the collective effort that contributed to their academic success.