Inspection and replenishment policies for systems with inventory record inaccuracy

For many companies, inventory record inaccuracy is a major obstacle to achieving operational excellence. In this paper, we consider an inventory system in which inventory records are inaccurate. The manager makes inventory inspection and replenishment decisions at the beginning of each period. There is a cost associated with each inspection. If an inspection is performed, inventory records are aligned with physical inventory. The objective is to develop a joint inspection and replenishment policy that minimizes total costs in a finite horizon. We prove that an inspection adjusted base–stock (IABS) policy is optimal for the single–period problem. In the finite–horizon problem, we show that the IABS policy is near optimal in a numerical study. Under this policy, the manager performs an inspection if the inventory recorded is less than a threshold level, and orders up to a base–stock level that depends on the number of periods since the last inspection. The prevalent approach to deal with inventory inaccuracy in practice is to implement cycle–count programs. Based on the structure of the IABS policy, we propose a new cycle–count policy with state–dependent base–stock levels (CCABS). We show that CCABS is almost as effective as the IABS policy. In addition, we provide guidelines for practitioners to design effective cycle–count programs by conducting sensitivity analyses on the IABS policy. Finally, by comparing the costs associated with these policies and several benchmark systems, we quantify the true value of accurate inventory information, which may be provided by radio–frequency identification (RFID) systems.