An agent-based model evaluating last-mile solutions for parcel delivery in an urban area

Abstract

E-commerce has grown rapidly worldwide over the past decade and thus more parcels are shipped into urban areas. The last mile delivery of these parcels into an urban area create big problems such as congestion and pollution. In order to tackle these problems new solutions have been found by either changing the operations of the system or using different vehicles. It is difficult to predict what will happen when these solutions are implemented in an urban area. Thus freight models are required to evaluate these last-mile solutions. This research used Agent-based modeling (ABM) to evaluate last-mile solutions for the delivery of parcels in a zero-emission urban area. The four last-mile solutions that were evaluated are vehicle type, collaboration levels, varying number of UCCs and varying parcel acceptance rates. The results show that electric delivery vans perform better than cargo bikes based on distance traveled, cost and number of vehicles required. On collaboration levels it shows that full collaboration can have a positive impact on distance travel traveled and cost. It is suggested to use two UCCs instead of five since having more UCCs does not decrease distance traveled and costs. The results also showed that when all parcels are accepted in the system on their first try it reduces cost, distance traveled by all vehicles and the total number of vehicles required. Local administrators can use these results when tackling the impact ever increasing growth of e-commerce and the parcels that need to be delivered.