Centralized Inventory Policies for 1-Warehouse/N-Retailers Systems Through a Simulation/Optimization Approach

In supply chain management, collaboration has become a key element to succeed. Operation strategies in which single components operate under local optimal conditions have been proven inadequate. Instead, the coordination of the entire supply chain is required. A 1-Warehouse/N-Retailers system typically operates with each retailer ordering products from the warehouse according to its own inventory policy. We approached this problem by allowing the warehouse to make centralized inventory policies, so that every participant in the supply chain can obtain benefits of economies of scale, while keeping or improving the service level to the final costumer. Our research question is: how to generate such those centralized policies? Centralized policies are only possible through a close collaboration of retailers and the warehouse. We used a mathematical optimization model to represent the system, and a combination of Monte Carlo simulation and a meta-heuristic optimization technique (Scatter Search) to solve the mathematical model. This combination is required due to the model complexity and stochastic nature of the system, i.e. the demand and some cost parameters can only be approximate by a probability distribution. Our results include the validation of the model to represent a real system. Additionally, we observed a reduction in overall operation costs of about 20%, while offering the same service level or higher for some costumers.  Our approach enables effective collaboration in the supply chain, which integrated with analytical tools can actually reduce the operation costs helping the companies to better compete in their markets.