Changes in the group
Maximilian Merkert moved to the TU Braunschweig where he was offered a junior professor position at the Institute of Mathematical Optimization. See this MathCoRe page for more details. Congratulations and good luck, Max!
Adrian Reimann and Christoph Plate joined our group as PhD students during the year 2021. After finishing his Master's thesis about machine learning methods applied to time series in acute lymphoblastic leukemia, Adrian started in April to work as a PhD student in our working group and continues to do research in the intersection of machine learning and optimization applied to medical applications. After finishing his Master's thesis on A linearization-based method for Partial Outer Convexification, Christoph started in December a PhD project in the intersection of optimal control, optimal experimental design, and machine learning. Welcome to the group!
We were successful in the DFG special focus program Machine Learning in Chemical Engineering. Together with the groups of Achim Kienle and Andreas Seidel-Morgenstern we launched our project Machine Learning for Power2Methanol Processes.
Completion of theses
Benjamin Peters defended his thesis Monomial Patterns in Polynomial Optimization. He introduced a generalization of various concepts in polynomial optimization that is at the same time elegant, efficient, and stimulating for future research. Additional reviewers of the thesis were Gennadiy Averkov and Christoph Helmberg. Clemens Zeile defended his thesis Combinatorial Integral Decompositions for Mixed-Integer Optimal Control. He defined the new state-of-the-art in constrained mixed-integer optimal control and could solve challenging applications in mobility and cardiology. Additional reviewers of the thesis were Christian Kirches and Sven Leyffer. Jennifer Uebbing defended her thesis Power-to-Methane Process Synthesis via Mixed Integer Nonlinear Programming. She introduced several models of different complexity for Power2X processes and developed novel optimization algorithms that leverage the different complexities. Additional reviewers of her thesis were Kai Sundmacher and Larry Biegler. Andreas Himmel defended his thesis Optimization-based Operation Strategy and Storage Design for Coupled Processes. He developed theory and novel methods to obtain a combined design of storage tanks and optimal operation of production systems using these storage elements. The thesis was advised and co-reviewed by Kai Sundmacher.
Also several Master theses were successfully defended: Susanne Anzer (Physiologie-basierte pharmakokinetische Modellierung von Methotrexat bei akuter lymphatischer Leukämie), Kilian Bock (Active Learning for maximizing model utility in material science on a small fractographic data set), Sarah Feldmann (née Ludwig, Der Two-Halfspace Closure), Jan Krause (Investigation of Options to handle 3D MRI data via convolutional neural networks - applications in knee osteoarthritis classification), Melina Licht (Modellbasierte Untersuchung von Cortisol in Zeitreihen), Christoph Plate (A linearization-based method for Partial Outer Convexification), Christopher Puest (Lösung eines Optimierungsproblems zur Erkennung von Vorhofflattern mittels Dekomposition in quadratische Teilprobleme), and Adrian Reimann (Machine Learning Ansätze für Zeitreihen bei Akuter Lymphatischer Leukämie).
Conferences and social events
Clemens, Mirko, and Sebastian participated in the virtual SIAM CSE 2021 conference with talks in the minisymposia on mixed-integer PDE constrained optimization. MathCoRe had its annual colloquium in Potsdam. In September, the work group spent a sunny day on the Elbe. Our christmas party took place in Il Capitello.
