Hybrid Quantum Computing Meets Use Cases

BMWK-Project »EniQmA« for the Systematic Development of Hybrid Quantum Computing Applications

Industrially relevant Quantum Computing (QC) applications are almost always hybrid – that is, classical systems are used in combination with quantum circuits. Variational algorithms play a central role in the NISQ (Noisy Intermediate-Scale Quantum) era to achieve quantum advantages. This makes the use of quantum computing more complex and mostly the applications arise in a very ad hoc or experimental way.

In the project »EniQmA« (Enabling Hybrid Quantum Applications) we are working together with partners from research and industry to systematize these hybrid approaches in a targeted way. We provide support in the cross-departmental project for the structured development of hybrid quantum applications through software, methods and tools. We help with the orchestration of classical software and quantum software. For this purpose, our EniQmA team creates a set of tools for the entire lifecycle of hybrid quantum applications.

Making Quantum Computing Applicable in Industrial Practice

In the project, this is done using concrete industrial use cases. We first consider large-scale use cases:

  • Industrial risk analysis
  • Anomaly detection in production processes

We also address climate and environmental research, transportation, insurance and finance, construction and geology, and production and automation in the broad scope of use cases.

Expertise Fraunhofer ITWM

In 2019, the Fraunhofer-Gesellschaft entered into a strategic partnership with IBM on the topic of quantum computing, which led to the establishment of a network of competence centers for quantum computing. We as Fraunhofer ITWM and Fraunhofer FOKUS are part of this network. On the part of our ITWM team, researchers from the divisions and departments »High Performance Computing (HPC)«, »Financial Mathematics«, and »Image Processing« contribute their expertise.

In particular, our knowledge of simulation, optimization, HPC, and machine learning is in demand. Our domain knowledge from our projects on the flexibilization of energy markets or risk management in the insurance industry, for example, also serve as a basis from previous research.

Structure of the Lifecycle of Hybrid Quantum Applications
© Fraunhofer ITWM
Structure of the Lifecycle of Hybrid Quantum Applications

The knowledge gained in the EniQmA project enables Fraunhofer to develop a holistic offering on quantum computing. In the future, interested companies can contact Fraunhofer and the quantum computing competence centers for comprehensive advice on the various aspects of hybrid QC algorithms and to advance their own industrial QC research projects.

EniQmA Software Environment

In the project we develop methodological approaches for hybrid classical / quantum algorithms. This includes, for example, decision support for the question of which part of the application is to be outsourced to quantum computers and which computer architectures, compilers, etc. are best suited for this. For the design of such a system we rely on already known solutions, which have been developed in the project »PlanQK«. In the PlanQK project, some of our project partners, such as Fraunhofer FOKUS, University of Stuttgart, Deutsche Bahn or Freie University Berlin, have collaborated and thus laid the foundation for our project »EniQmA«. Furthermore, we evaluate different classical optimizers for the hybrid variational algorithms as well as possible pre-optimizations.

Use Case: Optimizing Complex Processes in Production With Conspicuity Detection and QC

Very large amounts of data are generated in modern production processes. These are to be used to optimize process steps, to automate quality assurance more strongly and to dovetail it with the manufacturing process, to reduce construction deviations and to avoid nonconformities. In projects with various companies, we have already developed software solutions in the department »Financial Mathematics« that support the detection of anomalies. This makes it possible to find and examine different types of anomalies in large amounts of data – in the past mostly accounting data.

Conspicuity detection in production makes it possible to identify improvement measures for individual work steps or sub-processes at an early stage and thus optimize the production process. To do this, we analyze process data such as image data or time series to uncover deviations and weak points in production. This generates large volumes of data that have not usually been analyzed systematically to date. In the future, quantum algorithms will support this conspicuousness detection, for example, in order to identify errors in production processes. Currently, we are already using methods of machine learning (ML), artificial intelligence (AI) and statistical analysis. 

So-called Quantum Machine Learning (QML) promises to speed up algorithms and make them more reliable. First, we define requirements and interfaces for the application of hybrid quantum algorithms for abnormality detection in production processes. From this, we identify suitable subproblems for prototypical solution on QC, then implement the required algorithms and compare with ML or statistical methods.

Use Case: Advancing Aircraft Construction through Quantum-Assisted Composite Material Analysis

Modern aircraft construction relies on designing materials tailored to specific needs. Central to this is the utilization of composite materials which offer an ideal balance of strength, weight, and corrosion resistance. Such composites enhance the aircraft's performance and promote energy efficiency, aligning with the call for more sustainable aviation.

The process of designing these materials can be sped up using numerical simulations. With the aid of mathematical models and computer programs, engineers are empowered to test and optimize material structures virtually. This results in notable savings in both time and cost. However, due to the complexity of these simulations, computational costs can rise significantly, demanding efficient solutions.

While Quantum computing offers a promising solution, its direct application in numerical simulations can be challenging. Instead, our research emphasizes the development of quantum models based on DNS (Domain Name System) data to simulate effective material behavior efficiently. 

Four Primary Reasons Motivate this Approach: 

  1. Superior performance of quantum models over traditional networks with limited data
  2. Quantum models' robustness against disturbances
  3. Current quantum devices can execute these quantum builds.
  4. Avoiding the conversion challenge from classical to quantum data by pre-setting data embeddings

Our ambition lies in reshaping the aircraft construction sector. By coupling our theoretical advances with practical application, we envision a future marked by heightened precision, efficiency, and sustainability in aircraft design and operation.

Project Duration and Funding

The project is scheduled to run for three years from August 2022 to July 2025. EniQmA is funded by the German Federal Ministry of Economics and Climate Protection (BMWK) as part of the program »Quantum Computing – Applications for Business«. The total project volume is €9.3 million (of which €6.6 million in funding).