Project Details
Projekt Print View

Modeling of a hyperheuristic approach within an agent system to support operational planning for industrial product service systems in the production environment

Subject Area Production Systems, Operations Management, Quality Management and Factory Planning
Term from 2019 to 2023
Project identifier Deutsche Forschungsgemeinschaft (DFG) - Project number 424733996
 
Final Report Year 2024

Final Report Abstract

Industrial Product-Service Systems (IPS2) are transforming traditional business models by selling the availability or use of a product instead of the physical product itself. This offers manufacturing companies significant advantages, such as increased customer satisfaction or the generation of long-term, continuous revenue. At the same time, however, providers also face new challenges. One critical aspect is operative service delivery planning and scheduling, in which suitable resources such as technicians are allocated to specific orders, primarily maintenance activities. The complexity of this allocation is further increased by the variety of possibilities and the different requirements of locally distributed orders, such as deadlines and potential penalties. Inefficiencies in the planning process can not only lead to significant cost increases, but also affect customer trust. In view of the enormous solution space and the complexity of the planning task, an exact computer-aided solution in polynomial time cannot be realized, which makes the use of heuristic methods indispensable. To overcome this complexity and automate the planning process, in this project, a decision support system was developed for operative service delivery planning and scheduling in IPS2. To this end, the relevant actors in operative planning and scheduling were identified and analyzed using the Gaia method. Subsequently, the communication processes and interactions of the actors in the individual types of provision were examined in more detail with a focus on smart services. On this basis, an agent model was derived to enable the flexible adjustability of the decision support system. The agent model includes customers, orders, technicians, dispatchers, IPS2-Providers, and external suppliers as agents. For the automatic generation and optimization of plans regarding the costs to be expected from their execution, a hyperheuristic approach was modeled, which uses a Genetic Algorithm. This selects predefined low-level heuristics (LLHs) – problem-specific solution strategies – according to the principles of evolutionary theory and applies these to the problem so that the strengths of each individual LLH are utilized, and the weaknesses compensated for. The development of the decision support system resulted from the transfer of the hyperheuristic approach into the agent model, which was implemented in such a way that it can instantiated based on user input on technicians and orders and thus be adapted to specific business constraints. The developed system was first evaluated by creating optimized plans for randomly generated order data. Finally, a validation was carried out using real order data from a company, which proved the efficiency of the system.

Publications

 
 

Additional Information

Textvergrößerung und Kontrastanpassung