14:00-14:30 – Executable digital twins to measure the unmeasurable
Presented by Daniel De Gregoriis (Siemens), Frank Naets (FlandersMake@KULeuven)
Executable digital twins offer adaptable virtual representations synced with real systems, embedding them for seamless integration. They provide extensive insights like stress distributions and operational loads, replacing costly sensor setups with model-based estimation algorithms. While accurate measurements remain vital, digital twins offer substantial value in design validation, troubleshooting, and performance assessment across engineering stages, from development to production and exploitation.
14:30-15:00 – Using machine learning to predict your demand
Presented by Brecht Audenaert (WEngage), Dirk Van den Poel (FlandersMake@UGent)
This case study examines call center call forecasting at WEngage in Belgium, crucial for balancing service quality and cost-effective staffing. Leveraging machine learning, WEngage enhances forecasting, enabling informed decisions and productivity. Applicable beyond call centers, this approach aids demand prediction for various products or services. Due to diverse call patterns, a hybrid approach is vital for adaptability and accuracy. The study introduces two hybrid models, one based on a simple combination rule and the other using an advanced meta-heuristic approach. Both outperform individual models, with the advanced hybrid notably excelling, demonstrating superior consistency and accuracy in multi-step ahead call forecasting.
