Maintenance & after sales service supply chains

Chair: Matthieu van der Heijden (University of Twente)

Title: Operational Spare Parts Planning for Service Control Towers
Speaker: Sena Eruguz (Erasmus University Rotterdam)
Authors: A.S. Eruguz, R. Dekker

Abstract: In service control towers, real-time supply chain information (e.g., inventory, order progress, service level) are monitored. In this paper, we discuss how to use these information in operational spare parts planning under a finite-horizon service-level agreement (SLA). The SLA is a commitment on a certain service level (e.g., fill rate, availability, waiting time) over a specified time horizon. Penalty clauses apply if this commitment is not fulfilled. The objective is to optimize expediting decisions in order to minimize the expected total cost of holding inventory, expediting, and penalties for not meeting SLA.

Title: Reducing risks in spare parts service contracts with long-down constraints
Speaker: Douniel Lamghari-Idrissi (Eindhoven University of Technology & ASML)

Abstract: We are interested in spare parts service contracts for expensive assets. The contract performance is measured against the extreme long down (XLD) service constraint, which aims at limiting the number of long system downs. We consider a single item, single location stockpoint. In case of stock out, the demand for spare parts is satisfied in an alternative way. The expected costs of a spare parts service contract are evaluated in the contracting phase based on the expected spare parts usage, the contract duration and the targeted performance. During the contract execution phase, the true demand is observed and satisfied. The true demand over a finite horizon can differ from the expected demand due to the stochastic nature of failure processes and the total demand during the finite contract period. To this end, we introduce the usage-based contract, a contract defined over a predetermined number of usages as opposed to the traditional time-based contracts for which the contract duration is fixed in time. Using a Markov decision process, we prove that a state-dependent base stock policy is optimal. Using simulation, we compare our usage-based to the traditional time-based contract. The usage-based contract outperforms the time-based contract in dealing with the stochastic nature of failure processes. The savings increase as the reliability of the components increases. When the variation in demand rate is due to external factors, the usage-based contract results in significantly lower costs. The deviation of the total costs from the expected costs is also  lower when using a usage-based contract. The usage-based contract also delivers better performance to the asset user with less variation in the performance being delivered.

Title: Operational planning in service control towers – heuristics and case study
Speaker: Berry Gerrits (University of Twente)

Abstract: To operate spare parts networks, there are strategic (e.g., where to locate warehouses), tactical (e.g., the stock levels per part and per location), as well as day-to-day operational planning activities. In this talk we focus on the operational level and particularly on proactive interventions using a Service Control Tower (SCT). In a typical SCT, real-time data on the supply chain status are monitored, and alerts (triggers or exception messages) are generated in case of possible supply chain anomalies. Planners often analyze such anomalies manually, but decision support could be offered by presenting multiple intervention options with their estimated impact. We developed methods for trigger generation and heuristics to find good proactive interventions that improve SLA fulfillment and reduce the need for expensive reactive interventions, based on a case of a large IT-manufacturer.