Quantum Application Lab (QAL)

Problem: At Air France KLM, more than 7,000 employees work in shifts with different contract percentages, skills, and authorisation levels. Efficiently scheduling all these employees is a complex task. Traditionally, this is solved by designing basic and personal schedules, which reduces efficiency.

Solution: Quantum annealing is used to find the global minimum in a function, supporting and improving the scheduling processes. This technique is suitable for optimisation problems such as crew scheduling. By formulating the problem as a Quadratic Unconstrained Optimisation (QUBO) problem, we found solutions using quantum annealing, hybrid annealing, and simulated annealing.

Result: Quantum annealers achieve nearly optimal schedules, while hybrid solvers outperform both methods. In the future, quantum annealing can generate optimal schedules (with more variables) that comply with regulations, preferences, and operational needs more efficiently than traditional methods.

This work is supported by the Dutch National Growth Fund (NGF) as part of the Quantum Delta NL programme.

Problem: Bathymetry is the study of water depths, conducted by S[&]T using remote sensing data. This is important for the Dutch Ministry of Defence to gather useful intelligence. They use it, for example, to determine possible landing sites for amphibious vehicles or the presence of sandbanks. The depth of the ocean floor between 0 and 20 metres is derived from large amounts of multispectral satellite data using advanced machine learning algorithms. These data, consisting of 13 spectral bands with a temporal resolution of 5 days and a spatial resolution between 10 and 60 metres, result in an enormous amount of data that becomes unmanageable for classical methods.

Solution: Quantum neural networks can achieve high accuracies with small amounts of data because quantum machine learning algorithms can generalise well with little data. The challenge is loading large datasets onto a quantum processor. To make optimal use of quantum hardware, classical dimensionality reduction was applied to reduce the high spectral dimensionality of the available data. The data were then embedded in a variational quantum classifier to classify them into depth intervals.

Result: The quantum algorithm was executed on a 4-qubit simulator and tested on real multispectral data near the Mediterranean Sea. The classical results were reproduced with a quantum classifier. Measuring water depth more accurately with less data is beneficial for climate monitoring, resource exploration, and navigation.

Problem: Pulsar detection is a complex task in radio astronomy, involving the processing of vast amounts of data to locate these distant cosmic beacons. Together with ASTRON, we aim to refine the detection and analysis of pulsars, thereby enriching our cosmic knowledge. This is a massive task, partly due to undetectable cosmic phenomena.

Solution: This project leverages quantum computing to accelerate and improve the analysis of astronomical data. Two methods are central: a variational one-qubit classifier and a Variational Quantum Linear Solver (VQLS). The one-qubit classifier simplifies data categorisation, aiding in the rapid identification of potential pulsars. The VQLS, available on QAL's GitHub, offers a new approach to solving linear systems of equations for image reconstruction, a common challenge in pulsar data analysis, with improved efficiency and scalability through quantum computing.

Result: The application of quantum computing in pulsar detection can significantly reduce data analysis times and potentially reveal previously undetectable cosmic phenomena. The signal processing pipeline consists of many computationally intensive steps. The introduction of variational classifiers and VQLS in various parts of this pipeline lays the foundation for profound discoveries in the field.

Problem: Alliander is expanding the capacity of their network over the next 10 years by an amount comparable to the past 100 years. New technologies such as solar panels and electric vehicles add extra complexity to the electricity network. Together with QAL, Alliander is exploring the feasibility of quantum computing for optimising the robustness of energy networks in the event of cable failures, known as the 'N-1 problem'.

Solution: Three quantum computing solutions were investigated: amplitude amplification, quantum annealing, and Gaussian Boson Samplers. These methods utilise quantum parallelism and adiabatic evolution to efficiently monitor network reconfigurations and connectivity after cable failures.

Result: Tests showed that quantum computing has the potential to verify the robustness of large systems, which is unmanageable with current classical computers.

This work is supported by the Dutch National Growth Fund (NGF) as part of the Quantum Delta NL programme.

Problem: Toyota Motor Europe is collaborating with QAL on photocatalysis for water splitting, a crucial process for sustainable hydrogen production. Accurately simulating these reactions is complex and requires innovative computational methods.

Solution: The project uses an advanced quantum algorithm, the state-averaged orbital optimisation variational quantum eigensolver (SA-OOVQE), to simulate the energy interactions in photocatalysis. This algorithm enables accurate and efficient modelling of complex processes, providing deep insights into the interactions during photocatalysis.

Result: The use of SA-OOVQE in photocatalysis offers significant advantages. It improves the accuracy of simulations and provides deeper insights into reaction mechanisms, accelerating the research and development process. Simulations were conducted on a quantum simulator, and with further collaboration and hardware development, we expect to simulate increasingly larger systems.

Problem: Rapid economic growth, energy transition, and digitalisation in the Netherlands have led to increased electricity consumption. New offices, factories, residential areas, data centres, electric vehicles, and heat pumps demand more energy. Alliander seeks ways to improve the performance of their grid analysis and simulations, as traditional methods are no longer sufficient. Key tasks include performing power flow analysis and state estimation, which are computationally intensive.

Solution: Alliander uses quantum computing to accelerate their calculations. A crucial part is reordering matrices to improve the efficiency of calculations. This is done using a quantum algorithm that solves problems through Quadratic Unconstrained Binary Optimisation (QUBO). This approach has been tested on several practical datasets and has shown promising results.

Result: Accelerating these calculations helps quickly identify weak points in the grid and predict future energy needs. This is crucial for ensuring grid stability, maintenance, upgrades, and preventing outages, thereby guaranteeing a reliable and efficient electricity supply.

Problem: Genomic selection improves future generations by breeding animals or plants with desired traits. This process, which uses DNA markers and characteristics, promises higher yields, better quality, disease resistance, and sustainability. Together with researchers from Wageningen Livestock & Research (WLR), we are developing algorithms to tackle the computational challenges of large and complex datasets.

Solution: Researchers from QAL and WLR have developed randomised algorithms for singular value decompositions (SVD). These methods reduce the dimensionality of datasets, speed up calculations, and maintain accuracy. Quantum-inspired selection techniques focus on the top 5% with the highest genetic value, reducing inefficiency. Tools such as truncated SVD and Halko's algorithms significantly shorten computation times.

Result: This approach optimises genomic predictions, shortens the identification time of valuable individuals, and improves breeding programmes. This contributes to food security, biodiversity conservation, and more sustainable agriculture.

This work is supported by the Dutch National Growth Fund (NGF) as part of the Quantum Delta NL programme.

Problem: Organisations use portfolio analysis to examine how their business activities and investment or stock portfolios contribute to their results. For optimal risk and return distribution, scenarios are calculated and compared with objectives. As more variables are added or the analysis period is extended, these scenarios become more complex. This can lead to a large problem that is difficult to calculate or needs to be simplified, potentially resulting in suboptimal analysis.

Solution: TNO, together with Rabobank, investigated whether quantum computing can help optimise loan portfolios. In this case, financial objectives of the global loan portfolio in the agriculture sector were combined with environmental and sustainability goals. "We wanted not only a traditional analysis of revenue, profit, and return, but also to see the impact of our portfolio on the Paris climate goals," says Mischa Vos of Rabobank. The problem was reformulated as a Quadratic Unconstrained Binary Optimisation (QUBO) problem. Two QUBO formulations were presented, each with a different focus.

Result: Mischa Vos: “The result pleasantly surprised us. We quickly had an analysis that can serve as a basis for future decisions.” Frank Phillipson of TNO was also satisfied: “This practical example shows the potential of quantum computers.” Vos adds: “The board decides, but Rabobank wants to contribute to the 2035 climate goals. Our portfolio plays a crucial role in this. What do we invest in and what do we not?”

Read more about the results in this article and see how quantum computing can tackle complex optimisation problems in the financial sector. It also highlights the potential of quantum computing for more efficient and robust portfolio analyses. Or watch the video.