Model reduction for fork/join overhaul & repair systems with rotable inventory

Overhaul and repair services are an important segment of the remanufacturing industry, where an asset is disassembled, and component parts are repaired and then assembled to restore the asset to an "as new" condition. A key characteristic of such services is the wide use of rotable parts, i.e., using repaired parts from other assets for assembly, as opposed to waiting for the completion of part repairs from the original assets. However, it is difficult to evaluate the performance of such system. In this paper, and overhaul and repair system characterized by a fork/join structure with rotable inventory is studied. The effects of rotable inventory to system performance are analyzed, and an approximation method is then developed for evaluating key performance measures. By modelling the rotable inventory as a "negative" queue, a system with one overhaul center and one rotable repair shop is first reduced to a tandem queue. Numerical results demonstrate that by this model reduction, TAT can be efficiently estimated within ten percent loss of accuracy as compared with results obtained from hours of simulation. A system with one overhaul center and N rotable repair shops is then reduced to a fork/join system with multiple branches of tandem queues. With appropriate methods to analyze fork/join tandem systems, approximated performance measures can be evaluated.