A vehicle routing system supporting milk collection

This paper describes a computer-aided custom designed support system to reduce milk transportation costs. The model was developed in co-operation with Uruguay National Cooperative of Milk Producers (Conaprole). The vehicle routing model contains three inter-communicating parts: a) a Geographic Information System, b) external and internal interfaces, and c) a Vehicle Routing solver that includes heuristics inspired on the strategy of ‘cluster first, route second’. The distance and time matrices are calculated directly on the cartography. The first official digital map of Uruguayan roads was digitised according to the model requirements. As a result of the system implantation, large benefits are reported in haulage savings and in improved efficiency in milk collection management. The solution cost depends on the choice of clustering algorithm and on problem topology.