In the consumer packaged goods industry, usually a large number of final items are produced in a highly automated flow line production system. Florian Seeanner presents a mathematical model which is an improved version of the General Lot-Sizing and Scheduling Problem for Multiple production Stages (GLSPMS) to tackle the problem of shifting bottlenecks. Especially, a new meta-heuristic is developed which combines the principles of Variable Neighborhood Decomposition Search (VNDS) and Exchange. This VNDS+E heuristic is able to find very good solutions for the improved GLSPMS, including a real-world problem case. Finally, further model extensions, e.g., for scarce setup resources, are proposed.

Autorentext

Dr. Florian Seeanner wrote his dissertation under Prof. Dr. Herbert Meyr's supervision at the Chair of Production and Supply Chain Management at the Technische .Universität Darmstadt

Zusammenfassung
Due to a varying product demand (changing product mix) and different production speeds, bottlenecks may shift between the stages. In that case, a simultaneous lot-sizing and scheduling of these stages is recommendable. Hence, an improved version of the General Lot-Sizing and Scheduling Problem for Multiple production Stages (GLSPMS) was developed. Moreover, several reformulation techniques were applied to this model to solve it exactly. Besides, a new meta-heuristic which combines the principles of Variable Neighborhood Decomposition Search (VNDS) and Exchange was implemented to find good solutions, even for a real-world problem case. Finally, further model extensions, e.g., for scarce setup resources, were proposed.

Inhalt
Simultaneous Lot-Sizing and Scheduling for flowline production systems.- General Lot-Sizing and Scheduling Problem for Multiple production Stages.- Model reformulation & Heuristics.