CHAIR OF

OPERATIONS MANAGEMENT

In traditional scheduling research, machines are supposed to be only able to process one job at a time. However, in several real-world situations, machines can process several jobs in parallel up to a given capacity. One example is the growing of crops in greenhouses. The arising capacitated scheduling problems form a generalization of well-studied scheduling problems and have rarely been studied in the literature. We analyze the characteristics of these problems in various settings, such as flow shop or job shop. Tailored algorithms are developed to solve realistic problem instances considering multiple objectives.

View project in the research portal

Large manufacturing companies often manage a network of multiple factories, creating distributed scheduling problems. These problems involve assigning jobs to one of several distributed factories and sequencing the jobs within their designated factories. However, planning in distributed environments also requires the transport of jobs to factories. These transports are usually neglected in the existing planning approaches but can significantly impact the generated plans. We analyze the impact of transportation concerning classical scheduling objectives and environmental objectives such as emissions and energy consumption. Furthermore, inter-factory transports of intermediate goods are analyzed.

View project in the research portal

Scheduling is one of the most relevant planning tasks in operations management and describes the sequencing of jobs on machines. In practice, so-called hybrid flow shops (HFS) are often found, i.e., production lines in which several machines are available in the individual production stages. This means that, in addition to the sequence of the jobs, the jobs must also be assigned to the individual machines. Using lot streaming, i.e., the early transfer of sublots to subsequent production stages, can increase efficiency. However, quantifying the benefits of lot streaming is still necessary and tackled in this project. Because traditional planning approaches are usually unable to solve the complexity of this planning task, solution algorithms for this problem will be developed and evaluated. Problem properties are exploited to improve the approaches further. The developed algorithms can enable higher productivity and efficiency in industrial production in the future.

View project in the research portal

Existing powerful operations research methods enable the creation of highly efficient plans for deploying personnel and vehicles in rail transport. In the implementation, however, delays and breakdowns mean that the plans can often not be executed as intended. To meet this challenge, in the project VIPES, methods are developed to design schedules for traction units and shift schedules for train crews in such a way that they are efficient and reliable at the same time. This is to be made possible by intelligent interaction between optimization and simulation. Machine learning techniques are used to identify efficient and reliable solution structures, which will be used in the solution procedures.

View project in the research portal

Personal- und Fahrzeugkosten stellen sowohl im Personen- als auch Güterverkehr die wesentlichen Kostenbestandteile im Betrieb von Schienenverkehrsnetzen dar. Daher ein effizienter Einsatz dieser für Eisenbahnverkehrsunternehmen von zentraler Bedeutung. Gleichzeitig sind die relevanten Planungsprobleme sehr komplex und erfordern leistungsfähige Verfahren des Operations Research um praktische Instanzen lösen zu können. Das Projekt beschäftigt sich mit der Entwicklung dieser unter Einbezug praktischer Anforderungen. Insbesondere wird dabei die Schichtplanung von Zugbegleitern sowie das Locomotive Assignment Problem im Güterverkehr betrachtet.

View project in the research portal

Die Maschinenbelegungsplanung ist eine klassische Aufgabenstellung der Produktionsplanung - und steuerung. Meist kommen in der aktuellen Forschung dabei Metaheuristiken zum Einsatz. Im Rahmen dieses Forschungsprojektes werden insbesondere Erweiterungen des Hybrid Flowshop Problems betrachtet. Zum einen wird die Berücksichtigung des Energieverbrauchs analysiert, welches zu multi-kriteriellen Optimierungsproblemen führt. Ein weiterer Aspekt ist die Möglichkeit Aufträge in Unteraufträge aufzuteilen. Der Einfluss dieses sogenannten Lot Streamings auf die Fertigstellungszeiten und Zykluszeit wird im Rahmen dieses Projektes untersucht.

View project in the research portal