LIS: A Knowledge Graph-Based Line Information System

Abstract

In a manufacturing enterprise, like Bosch, answering business questions regarding production lines, involves different stakeholders. Production planning, product and production process development, quality management, and purchase have different views on the same entity “production line”. These different views are reflected in data residing in silos as Manufacturing Execution Systems (MES), Enterprise Resource Planning (ERP) systems as well as Master Data (MD) systems. To answer these questions, all data have to be integrated and semantically harmonized conciliating the different views in a uniform understanding of the domain. To fulfill these requirements in this specific domain, we present the Line Information System (LIS). LIS is a Knowledge Graph (KG)-based ecosystem capable of semantically integrating data from MES, ERP, and MD. LIS enables a 360 $$^{\circ }$$ view of manufacturing data for all stakeholders involved while resolving Semantic Interoperability Conflicts (SICs) in a scalable manner. Furthermore, as a part of the LIS ecosystem, we developed the LIS ontology, mappings, and a procedure to ensure the quality of the data in the KG. The LIS application comprises many functionalities to answer business questions that were not possible without LIS. LIS is currently in use in 12 Bosch plants semantically integrating data of more than 1.100 production lines, 16.000 physical machines, as well as more than 400 manufacturing processes. After the rollout of LIS, we performed a study with 21 colleagues. In general, the study showed that LIS in particular, and KG-based solutions in general, paves the way of exploiting the knowledge in manufacturing settings in a reusable and scalable way.