Performance of static priority rules for shared facilities in a remanufacturing shop with disassembly and reassembly

In this paper, we propose a stylized model of a basic remanufacturing shop that handles two remanufacturable products. Product A is comprised of two components A1 and A2, whereas product B is a single entity. After disassembly, component A1 is remanufactured at facility F1; component A2 and product B are remanufactured at facility F2. Both remanufacturing facilities have limited capacity, and are modeled as M/G/1 queues. First, we argue that, under the assumptions of our model, delaying a component to the shop after disassembly, which is a common release mechanism in actual shops, never improves system performance, measured in terms of total weighted average sojourn time (TWAST). Second, we show that the constrained optimal scheduling rule at facility F2 (constrained to simple non-preemptive static priority rules) that minimizes TWAST depends on the processing time characteristics of A1, A2, and B, and can only be found numerically, in general. Using an extensive numerical study based on a numerical approximation for product A's average sojourn time, we show, however, that using FCFS as a scheduling rule at F2 achieves similar TWAST performance, with an average increase of only 7.5%. We also perform a simulation study and show that a two-moment approximation for product A's average sojourn time performs well except for a narrow utilization band.