Minimising Total Costs of a Two-Echelon Multi-Depot Capacitated Vehicle Routing Problem (2E-MD-CVRP) that Describes the Utilisation of the Amsterdam City Canal Network for Last Mile Parcel Delivery

An increase in e-shopping and (last mile) parcel deliveries has contributed to a rapid growth of urban freight transportation. This generates major impacts on city sustainability and liveability. Current solutions for urban logistics concern road traffic, but multiple Dutch cities have an extensive range of city canals that could be used for freight transportation over water. It was investigated how the city canal network of Amsterdam can be utilised for last mile parcel delivery, and what the related effects are. A MILP formulation of a Two-Echelon, Multi-Depot, Capacitated Vehicle Routing Problem (2E-MD-CVRP) was developed. The model describes a network in which ships transport parcels to pre-determined satellite locations in the city centre, where the parcels are transferred to cargo e-bikes for the last mile of the delivery to the customer. The model was optimised by minimising the total costs, using the Genetic Algorithm (GA). The algorithm was able to find solutions but could not always stay within the constrained search space. Different possible network scenarios were evaluated, describing the consequences with respect to emissions, costs, and traffic flows. The results show promising economic, social, and environmental outcomes for a network with ships and cargo e-bikes instead of delivery vans. A daily and investment cost reduction of 16% and 36% respectively and a 𝐶𝑂2 emission reduction of 78.26% can be realised.