Project Duration: 01.01.2022 – 31.12.2024 | Project Partners: Chair of Production Informatics and Chair of Organic Computing at the University of Augsburg, ATR Software, BSH, Seele | Funded by: Bavarian State Ministry of Economic Affairs, Regional Development and Energy
Objective
ProsKI develops an AI-based assistance system for production planning and control (PPC) to handle disruptions and faults. The goal is to use machine learning to identify disruptions at an early stage, derive solution strategies and predict future disruptions – for shorter response times and more robust production control.
Results
The methodology combines data analysis, machine learning and a digital twin of the production environment.
Historical production data is analyzed to detect disruption patterns and generate forecasts. Explainable AI (XAI) ensures interpretable models that provide planners with comprehensible recommendations for action.
Industrial Benefit
ProsKI addresses the strong dependence on manual experiential knowledge in disruption management and rescheduling in manufacturing. The system automates the identification of deviations, supports solution strategies and integrates ML-based disruption forecasts into the production plan – for greater agility, transparency and acceptance among the workforce.
Significance
The project expands software4production's competencies in the field of AI-supported, resilient production planning. The insights gained on disruption management, forecasting models and Explainable AI feed into the further development of our Advanced Planning and Scheduling (APS) solution, making planning systems even more agile and fault-tolerant.
Publications
Publications related to this project can be found under Publications.
Are you interested in resilient PPC? Contact us – together we will discuss how ProsKI can make your production planning and control, and thus your value creation, more resilient.