Order and Stock Costs Optimization in an Automotive Spare Parts Wholesaler

Abstract

We address a demand forecast problem posed by Nors, a Portuguese group working in the automotive sector and transport solutions. The aim is to define a procedure that efficiently forecasts the references demand and to obtain an optimized procedure for their ordering methodology. Nors works with several suppliers, resulting in a portfolio of hundreds of thousands of different parts number. Each supplier has its own lead-time and order periodicity. On an annual basis, and for each supplier, the yearly budget for purchases is agreed in order to define possible quantity discounts and the profitability of the business. Sales distribution is highly scattered, as there are high, medium and low rotation sales values for different references. Since the number of references and their total value is very high, the stock value and operational costs are non-negligible factors concerning the company costs. Nors goal is to reduce these costs while maintaining a high service level. In this paper, we implement a mixed methodology for the forecasting problem, present an alternative for the safety stock value, and discuss an optimization model to decide which and how many parts number should be ordered in each period. Finally, we present a set of computational and real implementation results, obtained with Nors data regarding two suppliers.