Production–inventory planning in high-tech low-volume manufacturing supply chains

Abstract

This paper studies production–inventory planning in high-tech low-volume manufacturing supply chains, where long production lead times, complex network structures, multiple capacity constraints, as well as non-stationary demand and supply uncertainty, complicate production and safety stock placement decisions. The majority of academic approaches considers these problems under limiting assumptions about the system, while practitioners mostly rely on suboptimal heuristic integration of plans. We bridge the gap between stylized stochastic inventory models and industrial production planning practices by developing novel practice-driven multi-stage stochastic programming models. We validate our methodology in the uncapacitated single-product setting without lead time uncertainty by benchmarking strategic safety stock levels against an approach based on the seminal stochastic service model, which we tailor to accommodate the non-stationary demand. Additionally, we analyze the impact of capacity constraints and lead time uncertainty on safety stock requirements, and demonstrate how the performance of the benchmark solution regresses under these complexities. Finally, to support production–inventory decisions in view of the short life cycles of high-tech products, we adapt our modeling framework to new product introduction planning under modular product design, motivated by the setting of our industry partner, ASML. Following a fast-time-to-market strategy, we investigate optimal ramp-up and phase-out plans given interdependent demand and supply uncertainty due to risks in technology development. This study highlights the flexibility of our approach in addressing specific challenges encountered in real-world planning problems.