Integrating Neighbours into an Agent-Based Travel Demand Model to Analyse Success Rates of Parcel Deliveries

Abstract

The rapid growth of the e-commerce market leads us to expect a further increase in delivery vehicles in urban areas as well. This growth is expected to be accompanied by an increase in emissions while space becomes scarce. Meanwhile, people are adjusting their travel behaviour; therefore, the growing e-commerce market affects both last-mile delivery and private passenger traffic. Failed home deliveries are an important factor. They produce additional traffic by both ”Courier, Express and Parcel” (CEP) service providers and private passengers in the form of repeated delivery attempts or trips to pick up parcels. In this paper we apply an integrated agent-based model of last-mile deliveries and private travel demand. This allows for analysis of interactions between delivery and private passenger traffic and the status of the recipients during delivery. Furthermore, we present a neighbourship model to account for deliveries accepted by neighbours, which is crucial to reproduce realistic delivery success rates. We applied the presented model to the city of Karlsruhe, Germany, and simulated multiple delivery policy scenarios, which we compare to a static model without interactions between private and delivery agents. Our results show that the agent-based model produces more nuanced success rates with respect to different socio-demographic groups. Differentiating these groups is necessary when assessing measures that target specific groups and analysing effects of demographic changes. Also, we show the necessity of considering neighbours in such a model. This paper provides insight into the effects of e-commerce on a transport system and a framework to analyse policy measures or alternative delivery strategies.