Computational Framework for Multiagent Simulation of Freight Transport Activities

Abstract

It is widely recognized that micro-simulation and agent-based approaches can successfully be applied in transport policy analysis. However, logistic decisions and the complex relationships among freight actors make this a challenging task and a reason why the development of freight models is still behind the development of passenger models. In this paper, a multi-agent freight transport model is presented in which logistics decisions are separated into different roles: the shippers, which decide about shipment frequency, the transport service providers, which create transport chains, and carriers, which plan tours and schedule vehicles. All agent types can consolidate freight on their respective level and realize economies of scale. The lowest tier of the model, which contains individual freight vehicles, is integrated into the MATSim traffic simulation to create an integrated model for freight and passenger traffic. Changes in passenger demand, disturbances in the traffic system or policy measures can be picked up by freight drivers and propagated upwards to influence decisions on the levels of vehicle scheduling and transport chain building, and further on the level of shippers. As proof of concept, a scenario is set up with fictitious freight actors serving a set of customers. It is demonstrated that freight traffic can be simulated under different traffic conditions and policy measures.