A Data-Driven Concept and Realization for Engineering Change Management Decision Support.

In today's automotive industry, quickly reacting to engineering changes and providing market-driven products within a short period remains a competence for all Original Equipment Manufacturers (OEMs). Due to the complexity of products, changes to one component can lead to unexpected chain reactions in others. Besides, from creation to the approval of a change request can take weeks or even months without apparent reasons for the delays. To coordinate and control changes, companies established Engineering Change Management (ECM) processes. ECM processes can impact all determinants of competition of products: cost, quality, and time-to-market. Without proper management of Engineering Changes (ECs), negative impacts will happen. In the scope of this dissertation, a machine-learning based ECM decision support solution is developed which includes two main functions: change impacts prediction and lead time prediction. These functions support engineers to have an overview of change consequences in the early phase of the ECM process. The solution was evaluated based on the data from an automotive company and reached good performance. Therefore it was rated as beneficial to increase the efficiency, effectiveness, and quality of the existing ECM processes.