Last Time Buy and repair decisions for fast moving parts

Abstract

Spare part availability is essential for advanced capital goods with a long service period. Sourcing becomes challenging once the production of spare parts ceases, while the remaining service period is still long. In this paper, we focus on fast moving parts with repair of failed parts as an alternative supply option. We proceed from the methodology of Behfard et al. (2015) for slow movers, which assumes discrete demand distributions and therefore leads to excessive computation times for fast movers. We find that the use of continuous demand distributions requires significant modifications, both for the approximation of the performance indicators and for the optimization of the repair policy. We develop accurate heuristics to find the near-optimal Last Time Buy (LTB) quantity and the repair policy that we apply for two control policies: pull return - push repair, and push return - pull repair. We show that pull return - push repair is better to follow if return lead times are short and return costs are low. For long return lead times, we find that when the return cost exceeds 35%–40% of the part's value, push return - pull repair becomes more cost efficient. We also show that for relatively high demand of spare parts over the planning period (>300 for a 10 years planning period) the continuous model is a good approximation for the discrete model of Behfard et al. (2015). In addition, the computation time of our method is much lower then.