Data-based similarity assessment of engineering changes and manufacturing changes

The highly dynamic environment and the increasing complexity of products and production systems are forcing manufacturing companies to handle a high and further growing number of technical changes. Efficient change management is becoming an essential requirement for the long-term competitiveness of companies. One strategy in dealing with changes more efficiently is to learn from past changes and to use the gained knowledge. A necessary step for this strategy is the identification of similar past changes to enable further analysis. However, the identification of similar past changes represents a major challenge for change coordinators due to the variety and number of changes. Therefore, this work introduces an approach to assess the similarity of engineering and manufacturing changes based on structured as well as unstructured data extracted from IT systems used for the coordination of change management. It combines the methods of Natural Language Processing, clustering, and classification. The aim is to introduce an approach that meets industrial requirements and thus has the potential to support change management in practice. A data set of a medical technology company is used for a first industrial evaluation.