Exploration of Stochastic Lead Times Within the Aftermarket A case study at Volvo Service Market Logistics

Abstract

Within today’s global supply chains and competitive business environment, variations in demand and lead time contribute to that many organizations have difficulties in matching supply and demand. To manage uncertainties in demand and supply, many organizations use inventories and safety stocks, where the calculations often neglect variations in lead time, increasing the risk of having excessive or insufficient inventories. Furthermore, for the aftermarket which is characterized by high number of stock keeping units which have different demand patterns, there are limited literature available regarding how these variations can be considered within the inventory control system. The purpose of the thesis is therefore to investigate the lead time variability within an aftermarket setting as well as the potential impact of implementing stochastic lead times for determining safety stocks. The study has adopted a case study design with a mixed method approach and was conducted in collaboration with the Service Market Logistics department at Volvo Group. Several data collection methods have been applied, such as interviews with key stakeholders within the order process, reviewing of internal documents, an extensive literature review and the collection of historical data. The collected data were thereafter analyzed and followed by the development of a deterministic and stochastic model. The models were constructed to address the implications of incorporating stochastic lead times within an inventory control system. From the results of this study, it was found that lead time variabilities exist to different extents within different parts of the supply chain. To address these variabilities the main challenges are related to how spare parts with few shipments should be addressed as well as how a stochastic lead time parameter can be combined with a stochastic demand parameter when the demand during lead time is not normally distributed. Furthermore, by implementing a stochastic lead time variable the expected service level is assumed to increase when considering the whole lead time compared to using a deterministic lead time.