Optimization of Large Scale Parcel Distribution Systems by the Breeder Genetic Algorithm (BGA)

The routing and scheduling of vehicles and their crews is an area of increasing importance. In this paper, we describe a large scale vehicle scheduling and routing problem which is of great importance to parcel distribution companies. We devise a Breeder Genetic Algorithm capable of dealing with up to 10,000 transportation requests to be serviced by an inhomoge-neous fleet of vehicles within a 24 hour time interval. A transportation request is defined as the task to move a loaded container from one depot to another. Since the depots do not send out the same number of containers as they receive, the amount of empty containers available at the depots has to be balanced. The optimization task, therefore, is twofold: determine appropriate balance tours and find a low cost schedule for the fleet of vehicles.