Extracting Event Logs from Value Stream Simulation in Production Networks for Data Farming Based Strategic Network Design

Global enterprises operate complex production networks and aim to achieve an efficient value stream in strategic network design. However, network planners face the challenge of high uncertainties in planning parameters, particularly during the strategic planning phase, which can affect their decision-making process. To solve this problem, a data farming based planning approach is promising. Data farming utilizes a simulation model to execute a comprehensive experiment design that covers the uncertainties in planning parameters and generates a database from the results. This data is then used to support decision-making through process mining-based value stream analysis and evaluation, which helps to identify an efficient value stream. However, the lack of simulation models capable of extracting event logs, which are essential for process mining, is a critical issue. This work aims to establish event logs as an interface data format between simulation and process mining or advanced analytics. To achieve this, the contribution presents aggregated simulation techniques that are used to build a discrete event simulations (DES) model of a production network that is parametrizable in strategic network design and includes operational parameters such as inventory profiles. Additionally, a method to extract event logs from a discrete event simulation (DES) and an approach for handling the extracted data within a data farming approach are presented. Integrating event extraction and aggregated simulation modeling in DES helps to bridge the gap between simulation experiments and data-driven strategic network design. The applicability of the approach is demonstrated through the evaluation of an exemplary industrial example from the automotive industry.