Quantum computing in the supply chain

The experts in the Supply Chain Services working group combine their scientific expertise in quantum computing, mixed-integer optimisation and analytics with their domain expertise in supply chain management and pursue three approaches:

Solving application problems with quantum algorithms

The first research area focuses on the direct applicability of quantum algorithms to complex planning problems in the supply chain, in particular optimisation problems and forecasting time series. A common approach for this is so-called variational quantum algorithms (VQAs). These algorithms, which are based on physical concepts, are universally applicable, but must be re-parameterised for each specific problem. The aim of the research is to increase the applicability of VQAs through improved parameter optimisation. Good parameters are derived from characteristic parameters of the problem instance for special classes of problem types. As these parameters can be calculated efficiently, this method speeds up the search for good parameters. This improves the applicability of VQAs by making the parameter search more efficient.  Researchers benefit from the interdisciplinary interaction between physics and mathematics.

Combining the best of quantum algorithms and classical algorithms

Another research contribution is the development of hybrid algorithms. This involves combining classic optimisation methods with quantum algorithms such as VQAs. A major disadvantage of VQAs is their heuristic nature, i.e. the lack of statements about the quality of the solutions produced. On the other hand, they can quickly generate potentially good solution proposals. Embedding a VQA in a hybrid algorithm takes these two aspects into account: The higher-level classical method (for example ‘Branch and Bound’) works exactly, i.e. it finds an optimal solution or provides bounds on the quality of the solution found. The quickly available, heuristic solutions of VQAs can significantly accelerate such processes.

In this context, the researchers are utilising their experience in the development and use of classical algorithms with the aim of identifying places within a classical algorithm where it can benefit from quantum subroutines (this should be explained). The selective use of limited quantum resources enables them to be utilised more efficiently than in pure quantum algorithms.

Better solutions through optimised compilation

The third research focus is the investigation of combinatorial optimisation problems that arise when compiling quantum algorithms.  In order to be able to execute a quantum algorithm, it must first be translated into the language of the quantum computer to be used - similar to classical algorithms. This involves new types of combinatorial optimisation problems. The quality of the solutions to these problems is decisive for the error-proneness of the executed algorithm. Therefore, the development of improved solution methods for these problems directly increases the usability of current NISQ hardware.

Thanks to many years of experience in the development of solution methods for combinatorial optimisation problems, including in the supply chain context, the researchers can adapt known algorithms to the new problems and develop new methods at the same time. Good compilation can thus contribute significantly to the efficient utilisation of available quantum resources.