SAMS Subject Classification: 25, 26

Globalisation and the growth of e-commerce have led to retail companies having to manage and control larger and larger numbers of stock keeping units (SKUs). The success of any retail company depends on how well it can satisfy demand while remaining financially viable. Inventory management systems are typically aimed at balancing the conflicting objectives of achieving good customer service levels and minimising inventory and operating costs. As the number of SKUs a company holds increases, however, so too does the complexity of this balancing problem. The main decisions in respect of SKU inventory management in a retail warehouse, which affect this balance, are related to (1) which SKUs need to be replenished, (2) when to place replenishment orders for these SKUs, and (3) the appropriate volumes of SKUs to include in these orders.

In this presentation, a generic framework is proposed for the development of decision support systems in the context of SKU inventory replenishment in a retail warehouse. This includes a time series forecasting approach which utilises supplementary data on customer demand and generates forecasts of demand distributions. It also includes an inventory replenishment model according to which orders for SKUs may be placed at discrete, equi-temporal points in time. The objective is to batch SKU replenishment orders together, while accounting for lead times, minimum order quantities and backlogged orders. Model performance is evaluated on the key performance indicators of customer service level and cost, which form the typical trade-off in inventory management. The model is solved exactly. The functionality of an instantiation of this framework is presented in the context of a case study involving real demand data in the South Africa retail sector.