Forming an Inventory Control Policy in a Stochastic Environment: A Case Study at Sandvik Crushing and Screening

Abstract: Background: For most companies that handle large material flows there are buffers between different stages in the supply chain, often referred to as inventory. The main reasons for having these inventories are to achieve economies of scale when handling material flows and to buffer against uncertainties when matching demand and supply. Hence, the main objective in inventory control is to decide when and how much to order and by doing so, balance conflicting goals. These goals could be to reduce operational costs and to reduce stock on hand while achieving a certain service level to the customers. Purpose: Forming an inventory control policy for a new warehouse which is based on the demand from both the production of new crushers and from the aftermarket. Research questions: (1) How to form an inventory control policy for a new warehouse? a) How to optimize the parameters in this policy to minimize the ordering costs and the inventory holding costs? (2) How does the lead time variability affect these costs? Methodology: In this study, a framework presented by Yin (2009) and Hillier & Liebermann (2010) has been used to construct the research design. The initial step was to specify the problem at hand before the gathering of data was initiated which also included a literature study. Based on the gathered information, an initial analysis of the current system was performed before the mathematical model which represents the new inventory control policy was created. A statistical analysis of the lead times was conducted to provide input to the model. Finally, the constructed model was evaluated in a sensitivity analysis. Conclusion: Based on the analysis of sales data, a compound Poisson process was chosen as an appropriate demand model. From the theoretical foundation and current practices at the case company, an (R, Q) policy with periodic review was chosen as the new inventory control policy. Since the case company currently experience variability in the lead times, this uncertainty was incorporated in the new policy. The study showed that in the future state, where the case company would operate with a shared warehouse with 95% fill rate and take stochastic lead times into consideration, the holding costs would increase while the ordering costs would be reduced. When analyzing the effect of lowering the lead time variability in the new policy, it was shown that the holding costs were reduced while the ordering costs increased. Finally, it was concluded that in the best-case scenario where lead times are assumed constant, the total costs in the future state would be reduced compared to the current system.