Full Load Vehicle Routing with Multiple Depots: A New, Network Flow Based Algorithm

This paper considers the problem of scheduling a fleet of identical vehicles to execute a set of freight orders in truckload quantities. We propose an algorithm that starts from an optimal solution to a network flow model which considers all data in the problem in parallel including unloaded travel costs, freight orders, depots, availability of vehicles and the vehicle receiving capacity at each depot. The network solution is then iteratively modified through heuristic operations and the routing of vehicles is determined based on the modified solution. A novel feature of the algorithm is the option of resolving the network model for scheduling the remaining vehicles.