Smart Scheduling Recommender System
The FHNW Institute of Business Engineering is developing, in collaboration with industry partners, a Smart Scheduling Recommender System for process-centric production planning in medium sized enterprises.
Proof-of-concept for the powerful combination of process mining and process simulation algorithms in a Smart Scheduling Recommender System to optimize the planning process and to achieve significantly shorter throughput times with the application of a digital twin.
Switzerland's mechanical and electrical engineering industry (MEM industry) generate 7.1% (2018) of Switzerland’s GDP, thus occupying a key position in the country's economy. With exports accounting for 79% of their output, the MEM industry are highly export-oriented.
In the Swiss MEM industry, usually ERP modules for scheduling processes based on heuristics are applied in the production planning, which, often without load-dependent order control functionality, are designed to achieve high machine capacity utilization and to keep the safety stocks for final products. In connection with fast-track orders and open production orders (work in progress, overdue production orders), throughput times are constantly increasing due to long order queues at a flat and maximum achievable machine capacity utilization level. The result is cumulative fluctuating forecasts and subsequent troubleshooting, which can lead to situations that limit customer satisfaction and competitiveness.
The aim of the research project is to show how the throughput times in production can be significantly reduced and thus increased flexibility and responsiveness as well as a higher readiness to deliver for medium-sized companies in the MEM industry. Existing heuristic planning algorithms are expanded by a Smart Scheduling Recommender System, which is based on an intelligent digital twin and on real production processes and machine capacities. For the first time, process mining and process simulation algorithms are combined in a real production planning environment.
Project information | |
Implementation | |
Duration | 18 months |
Funding | |
Project management |