Industrial use cases by PlanQK & Bayern Innovativ

Duty rosters in healthcare are still created manually in many cases today, even though an almost unmanage-able number of factors such as qualification profiles, resource requirements, legal regulations and personal preferences have to be combined. Especially during the Corona crisis it became obvious how great the need for automated solutions is in order to react quickly to short-term changes and to be able to relieve em-ployees in an overloaded system in the best possible way. As part of the PlanQK project, Planerio Quantum Computing is researching approaches that can solve this problem efficiently and optimally.

Nicolas Butterwegge has a background in engineering, innovation and strategy. Previously he worked as an Innovation and Incubation Manager at Siemens Healthineers where he is currently working as an IT Strategy Manager. Before that he gained experience at Porsche, Bosch and Brose.

The COVID-19 pandemic has posed significant chal-lenges to healthcare systems around the world. Re-searchers around the world have developed machine learning algorithms based on radiographs and CT scans to detect and predict COVID-19 infection. How-ever, existing models have been criticized and further research was needed to improve the quality and clini-cal utility of these models. Quantum-based machine learning is believed to have the potential to surpass the performance and computational complexity of classical machine learning methods.

The goal of our research was to perform quantum-based AI analysis and automatic classification of lung CT images. A pipeline was used, where the segmenta-tion of the lung as well as the cropping to the relevant structures was done in a classical way and then a QC classifier performed the classification. The results of the analysis are automatically transferred to a struc-tured COVID-19 report template of Smart Reporting, resulting in a medical report.

Responding to a recommendation of the International Maritime Organization asking operators of ocean-going vessels to reduce speed such that the RTA will be reached Just-In-Time, FCE works on speed profiling algorithms, both classical and quantum oriented. For a fixed trajectory precomputed by traditional navigation tools, the task at hand is the generation of an optimal speed profile along the voyage, including environmen-tal conditions. Classically this requirement leads to a Dynamic Optimization  problem as expressed through the Stochastic Bellman Equation. From the Quantum Computing point of view the problem can be expressed as a so called QUBO (=Quadratic Unconstrained Op-timization Problem), as long as the fuel cost is no more than quadratic as a function of speed. However, rather than computing the speed profile as a sequence of individual values, it can also be expressed as a super-position of standard functions, whereby the only varia-bles to be found are the weight coefficients.

Quantum Computing is obviously a complex technology nevertheless it promises the next big thing. Many of us are surprised from ChatGPT and its risks and opportunities. To be prepared for the next big thing Deutsche Bahn started its journey on quantum in 2020 becoming a partner of the PlanQK consortium to implement a railway use case on an annealer. Deutsche Bahn build up groups for quantum computing as well as for quantum security (PQC, QKD) and is strongly integrated in the European ecosystem. In the speech I would like to share the experience we made to establish a completely new technology in a large enterprise and give you some results we gained so far on the quantum marathon.

Integrating quantum computing into existing software solutions sounds like a lot of headache – new installation of software and retraining of staff are huge upfront costs in addition to the recurring cost of using new software. The potentials of quantum applications make this equation even more complex with the additional factors of new hardware and software partners. We have taken an example from the parcel delivery industry to show how quantum applications can in the future be more easily integrated into existing process and software landscapes whilst still making use of all the required advantages.

No one knows when quantum computers will become relevant for industrial applications. But it is clear that once it happens, it is going to happen quickly. Especially in industries where most of a company's competitive advantage stems from solving computationally expensive problems, it is thus important to prepare for the future by analyzing the relevant computation tasks and identifying the impact of future technologies, all while the right experts are still available. Also, such an investment does not need to only yield a return in the uncertain future: Gaining new perspectives on a company's processes can improve their performance immediately.