Quantum methods and Benchmarks for Resource Allocation

The challenge of optimizing complex industrial processes, such as microchip fabrication and automobile manufacturing, demands the analysis of vast amounts of data to improve production allocation within capital-intensive facilities. Even minor enhancements to the solution quality of combinatorial optimization problems can lead to significant economic impacts for companies in these sectors.

For instance, Infineon Technologies AG, a leading semiconductor manufacturer, collects millions of data points daily from various sources to optimize its microchip production process. On the other hand, Volkswagen AG, a world-leading car manufacturer, grapples with the complex challenges arising from the transition to electric vehicles and mobility services. In both cases, effective optimization can yield substantial economic benefits for these companies. These combinatorial optimization problems are typically NP-hard, implying no fully general efficient algorithmic solutions exist. Practitioners resort to domain-specific heuristics and, more recently, machine learning algorithms to discover optimal solutions under the constraints of time and computational power. Quantum computers possess the capability to surpass classical computational processes for select applications, suggesting their potential in combinatorial optimization. Determining the feasibility of quantum algorithms hinges on multiple factors, such as problem specifics, hardware capabilities, and competing classical methods. Consequently, a comprehensive analysis necessitates robust collaboration between experts in quantum information, classical deterministic algorithms, machine learning, software engineering, and industry stakeholders. The goal of QuBRA is to assemble a diverse consortium with the aim of quantitatively outlining the practical quantum advantage in combinatorial optimization—a question of great relevance to industrial applications that currently lacks a clear understanding.