Optimization of Dynamic Repair Priorities in Spare Parts Supply Systems

We consider a single-location spare part supply system for repairable spare parts where all parts are repaired at the same single-server facility. A failed part is replaced immediately by a working one from the stock (if available) and is sent to repair. After the repair process, the part goes to the stock as good as new. This system is also equivalent to a make-to-stock system, where instead of broken parts orders for new made items are coming into the system. If a replacement item is not available due to long processing times or low stock levels, it is backordered and penalty costs occur. The repair shop can process different types of parts and can choose which part to repair according to the current stock levels or the current numbers of backorders in the system. We show how to optimize the processing priorities for different system states, and also present several heuristics for integral optimization of the stock levels and of the processing priories. The obtained results are compared to existing results for systems where processing priorities depend only on parts properties (and not on the system state) and to results for systems where parts are repaired according to First-Come-First-Served rule.

Remember that this is a lunch meeting. For ordering the right amount of sandwiches, coffee, tea and milk, do let us know before Wednesday October 18th whether you will be present. We hope to welcome all of you.