Order advancement in a multi-item two-echelon system: Theory and case study

Motivated by a case at a large retail chain in the Netherlands, we study short-term order advancement in a multi-item, two-echelon system consisting of a distribution centre (DC) and many stores. Application of traditional inventory replenishment rules may lead to a strongly fluctuating workload for order picking, and thus capacity problems, at the DC. By advancing replenishments we can balance the workload at the DC without negative impact on the service levels towards the final customer, trading off backroom usage and handling of incoming orders under limited shelf space at the individual stores. We develop a performance evaluation method and various heuristics to find good order advancement solutions. We select the heuristic with the best trade-off between cost performance and computation time for large problem instances (i.e., hundreds of stores and thousands of stock keeping units). Application of this heuristic to case data from the retail chain shows around 10 % overall cost reduction. In particular, the capacity shortage at the DC – a major issue in our case study – is reduced by almost 67 % at the expense of more peaks in handling workload at the stores (42 % increase in order lines exceeding store capacity), and an increase of 7.6 % in backroom usage. Sensitivity analysis shows that a planning horizon of about two weeks performs best, and that the cost reduction potential is heavily influenced by the available shelf space.