Last update: August 2026
Real applications come with constraints, structure, data, performance requirements, and strong classical methods that any quantum approach has to compete with. They therefore provide a demanding setting in which to study what quantum computing can realistically offer.
In my group, we investigate quantum algorithms through scientific and industrial applications, with a particular focus on combinatorial optimization and engineering problems. We study how such problems can be represented for quantum computation, how quantum and classical methods can be combined, and how algorithmic choices interact with the capabilities of the underlying quantum system.
A practical application usually needs substantial reformulation before it can be addressed by a quantum algorithm.
The original problem has to be translated into a representation that captures its objectives and constraints while remaining suitable for the selected quantum computing approach. This formulation step can have a major influence on the resulting computational problem.
A recurring focus of our work is quantum optimization, including Quadratic Unconstrained Binary Optimization (QUBO), quantum annealing, and variational algorithms such as QAOA. We study how constrained problems can be encoded efficiently, how relevant problem structure can be preserved, and how different formulations affect the resulting optimization landscape.
The formulation therefore becomes part of the algorithm design itself and strongly influences what the quantum method is asked to solve.
Many practically relevant optimization problems are dominated by constraints. Feasible solutions may have to satisfy assignment rules, cardinality restrictions, physical design requirements, or more complex relationships between decision variables.
We develop methods that incorporate this structure directly into the optimization process. Our work includes penalty-free formulations, systematic construction and simplification of QUBOs, constraint-preserving mixer Hamiltonians, and warm-starting approaches that use information from classical optimization.
Hybrid quantum-classical methods are particularly important in this setting. Classical methods can provide good initial solutions, simplify problem instances, or handle parts of the optimization process efficiently, while quantum methods can be applied to selected components.
We use these approaches to study which parts of a realistic optimization pipeline may benefit from quantum computation and how quantum resources can be integrated effectively with classical algorithms.
Real engineering problems provide concrete benchmarks for quantum algorithms because their constraints and performance requirements cannot be ignored.
One example is our work on composite laminate design, where suitable stacking sequences have to be found for lightweight composite structures. Together with colleagues in aerospace engineering and numerical analysis, we developed methods combining quantum computing and tensor networks for this problem. This work also formed part of our contribution to the Airbus-BMW Quantum Computing Challenge, where our team received the Golden App Award.
Applications such as this allow us to study questions that are difficult to capture with synthetic benchmarks alone. How large are relevant problem instances? Which constraints dominate the formulation? What solution quality is required? How does the method compare with strong classical approaches? And what quantum resources would be required for meaningful instances?
We use such applications as full-stack benchmarks. Their computational requirements provide input for algorithm design, compilation, and architectural evaluation.
The central question is where the structure of an application and the capabilities of quantum computation provide a meaningful match.
Answering this requires realistic problem formulations, strong classical baselines, reproducible evaluation, and explicit resource estimates. These elements allow us to assess both the potential and the limitations of a quantum approach.
Applications also provide requirements for the rest of the computing stack. The size and structure of a problem influence algorithm design, compilation overhead, reliability requirements, and ultimately the architecture needed to execute it.
In this way, application research helps us understand what a quantum computer has to provide to become useful for a real computational problem.
Constrained Quantum Optimization via Iterative Warm-Start XY-Mixers. New Journal of Physics, 2026.
Penalty-free approach to accelerating constrained quantum optimization. Physical Review A, 2025.
Quantum computing and tensor networks for laminate design: A novel approach to stacking sequence retrieval. Computer Methods in Applied Mechanics and Engineering, 2024.
SATQUBOLIB: A Python Framework for Creating and Benchmarking (Max-) 3SAT QUBOs. International Conference on Innovations for Community Services, 2024.
Strategic portfolio optimization using simulated, digital, and quantum annealing. Applied Sciences, 2022.
A hybrid solution method for the capacitated vehicle routing problem using a quantum annealer. Frontiers in ICT, 2019.
SATQUBOLIB: Python framework for creating, simplifying, and benchmarking QUBO formulations for SAT and Max-SAT problems.
[GitHub] [PyPI] [Publication]
Laminate Design Code & Data: Research code and datasets accompanying our quantum-computing and tensor-network approach to stacking-sequence retrieval in composite laminate design.
[Code] [Dataset] [Publication]