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

Abstract

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 predetermined 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.

Urban Logistics through River: A Two-Echelon Distribution Model

“…Although vessels are used for long distances, few studies in the scientific community consider making parcel distribution through the city's waterways with small boats. However, some studies in Amsterdam proved that this initiative could reduce congestion in the city center, satisfying the delivery needs of retailers without interfering with other fluvial activities [38,39]. Consequently, we want to address the 2E-VRP in two different scenarios that consider multimodal transportation to first determine the allocation of customers to the satellites, then to find a set of routes for the first and second echelon, so electric cargo bikes (e-cargo bikes) perform the last-mile delivery.…”