Purposeful decision-support with Digital Shadows: a model catalog architecture and its implementation for production scheduling in the injection molding domain

Industry 4.0 promises more efficient decision support systems by enabling the retrieval of data from various and heterogeneous sources and processing them in appropriate models. Consequently, Digital Shadows are introduced to automate the identification of models for specific issues and retrieve necessary input data automatically, streamlining decision-making processes. However, the absence of an architecture containing a model repository, information on model processing, required model input data, and their respective source systems poses challenges to implement this automated process. To address this, we propose a model catalog architecture as the cornerstone of automated model processing. Therefore, we ascertain the fundamental architectural requirements by delineating the essential interfaces, gates, and databases involved. These requirements are then translated into a modular and scalable architecture, facilitating the reuse of repetitive files, such as a translator acting as a middleware between different source systems. The architecture enables seamless integration of new models, provided they adhere to standardised definitions. Additionally, we transfer the architecture into a suitable framework. To validate our concept, we implement an exemplary model catalog that demonstrates production scheduling in an injection molding shop floor, completed with a web-based interface and integration of an enterprise resource planning system for real-time data access.