Concept for the Evaluation and Prioritization of Machine Learning Use Cases in Industrial Production

In the course of the advancing digitalization of industrial production, many enterprises have already laid the foundations for a more comprehensive end-to-end recording and accessibility of production related data. Machine learning (ML), implemented in specific industrial use cases, offers the possibility of automated analysis of these large volumes of data with considerably reduced manual effort. In industrial practice, however, the selection of use cases with an economic and long-lasting strategic impact poses challenges, since much of the decision-relevant information of individual use cases is mostly discovered during the actual implementation phase. Additionally, as the datasets required for a successful application are often not sufficiently known prior to this phase, a previous assessment regarding the data basis for individual use cases is also needed. To address these challenges, this paper presents a concept constructed in the research process for applied sciences according to ULRICH for a-priori evaluation and prioritization of use cases for machine learning in industrial production. In particular, the potential benefits, implementation efforts, and the technical feasibility are considered as evaluation dimensions.