This course is taught in English. This seminar was last offered in the summer term 2022.
Over the recent years quantum annealing inspired approaches have risen in popularity for solving business-related optimization problems. The list of firms developing special purpose hardware for these problems includes big names such as Fujitsu, D-Wave and Google, with even the Deutsche Bahn investigating their applicability. Fundamentally these techniques solve so-called QUBOs, quadratic unconstrained binary optimization problems. In this seminar we will investigate properties of QUBOs, discuss solution techniques and explore how they relate to various practical and academic combinatorial optimization questions.
We are looking for enthusiastic students who have an interest in optimization and integer programming algorithms. The papers to be covered are sometimes technical, and a basic familiarity with operations research and complexity theory is advantageous.
In total 10-15 topics will be covered, according to the number of participants. Each student will receive one topic, which will involve working through a paper or problem, giving a 30 min presentation about it, and writing a 5-page summary. This seminar also offers the exciting opportunity to model optimization problems, and solve them using Fujitsu's Digital Annealer, which is a state-of-the-art near-quantum computing machine. Given applicability to the topic an illustrative implementation will thus also be expected.
There will be four sessions in total. First a meeting on the 8 April at 3pm to discuss and assign topics, then two gatherings after three weeks respectively, and one final day on which all presentations will be held. We are targeting at most 10-15 participants. The remaining exact dates are still to be determined.
The following is a rough outline of topics that will be sketched and assigned in the first meeting: