Stochastic leadtimes in two‐level distribution‐type networks

Leadtime uncertainty occurs in both manufacturing and distribution systems and can cause difficult coordination problems. We analyze a simple three‐location, two‐level distribution‐type system with stochastic leadtimes with the objective of determining planned leadtimes which minimize the sum of expected inventory‐holding and tardiness costs. In a manufacturing context, the system can be viewed as one in which common processing or procurement is done first, whereupon another manufacturing stage differentiates this common product. Within a distribution framework, the system is one in which material is transported to a central facility (such as a regional warehouse) and subsequently transported to smaller local distributors or retailers. We investigate two heuristic policies which are simple adjustments to optimal solutions for serial systems resulting from a decoupling of the distribution network. We also report computational experience which indicates that the optimal “location” and quantity of safety time depends largely upon the relationship among the due dates and average leadtimes for the final stages of processing or delivery.