A Simulation-Based Approach to Multi-Item Multi-Echelon Spare Parts Inventory System Optimization with Variable Warehouse Roles

An optimized supply chain is essential for the success of large-scale industries. In order to meet the high availability level requirements, an efficient inventory system is crucial to reduce downtime when a failure event occurs. In a multi-echelon spare parts inventory system, each warehouse within the system may operate as a hub or a spoke. A hub is a warehouse that fulfills the demands of other warehouses, while a spoke is a warehouse that fulfills the demands of final customers. When the inventory system deals with multiple items, a fixed role (hub or spoke) is commonly assigned to each warehouse. However, this limitation may lead to sub-optimal solutions. If the optimization model allows each warehouse to have a different role for different items, a new degree of freedom is included and more efficient solutions can be found. In this paper, we propose a simulation-based optimization model to define the configuration of a multi-echelon spare parts inventory system of multiple items. The goal is to minimize total inventory costs, subject to a fill rate constraint. We relax the assumption that warehouses have a fixed role for all the items. Two algorithms are used to evaluate the model: the Teaching-Learning Based Optimization (TLBO) and the Simulated Annealing (SA) algorithms. A case study based on a spare parts inventory system of an aircraft manufacturer is used to compare the performance of the proposed model with the performance obtained considered fixed warehouse roles. The results showed that the proposed model provided a reduction of 6:8% in total cost, without violating the fill rate constraint.