Optimization of customer orders routing in a collaborative distribution network

This paper presents a sequential approach for the assessment of a multi-layered distribution network from a cluster of collaborating suppliers to a large set of customers. The transportation network includes three segments: suppliers routes from suppliers to a consolidation and distribution center, full truckload routes toward regional distribution centers, and less-than-truckload distribution toward final customers. In every shipping date, the optimization problem consists of assigning customers to regional distribution centers and determining the routes of vehicles through the whole distribution network. This problem is first modeled as a Mixed Integer Linear Problem (MILP). Then, we propose to decompose it into three smaller MILPs that are solved sequentially in order to quickly provide a good approximate solution. The experiments on real data show that the decomposition method provides near optimal solutions within a few minutes while the original model would require hours of calculation.