Inventory control in a centralized distribution network using genetic algorithms: A case study

• We use GA to obtain the optimal inventory levels in a real life distribution system.
• We investigate an inventory distribution system under stochastic and varying demand.
• We present a framework to obtain the cost parameters to be used in the model.
• Proposed system delivers substantial cost savings compared to the current system.

This paper presents a case study to determine the optimal inventory levels in a spare parts distribution system. We develop a solution based on a Genetic Algorithm (GA) for an effective management of the distribution network of a Turkish automotive manufacturer under centralized control. We provide a specific approach to address the two-echelon inventory control problem in its combinatorial and sequential behavior, dealing with a large number of specific properties that are considered in practice. Findings of the case study reveal that the use of the proposed inventory control system may provide substantial cost savings to the case company. We finally draw conclusions from the case study on the company’s operational practices and illustrate opportunities for improvement.