Ontology-driven data input for optimization

A major challenge when optimizing production facilities, whether in planning processes or with running facilities, is to describe the machines' initial state and to identify relevant optimization parameters. These factors have crucial influences on the optimization results. For conducting a simulation, relevant input data, representing starting conditions and influencing factors, have to be identified and acquired likewise. However, with growing complexity of systems, data collection is increasingly demanding and time-consuming. This task necessitates both knowledge about the process and the targets of improvement as well as a profound understanding of the system to be optimized. Actually, experts mostly have a thorough knowledge about either the optimization methodology or the production system. This leads to inefficiency when setting up the optimization or simulation base. In this paper, we present the approach of ontology-driven data collection for optimization tasks. Using a meta-ontology of a production facility, the user is guided through the process of data gathering. Depending on the improvement task, the ontology provides relevant parameters that have to be inquired. Thereby we provide a methodology to identify the entirety of input data for the simulation. The process is also applicable to facilities being planned. Thus the knowledge base can be used to support improvement tasks within the digital factory.