Complexities in moving from commodity to vehicular flows.

Abstract

Paper presented at the 32nd Annual Southern African Transport Conference 8-11 July 2013 "Transport and Sustainable Infrastructure", CSIR International Convention Centre, Pretoria, South Africa.

Two main schools of thought exist in the modelling of commercial vehicle movement. Firstly, the top down commodity flow models start at an aggregate level using metrics such as Gross Domestic Product (GDP) to derive the origin-destination flows of different commodities. The vehicle flows are then inferred using, amongst other things, traffic counts. The downside is that vehicle activity chains are disregarded, and load factors are overly simplified. Secondly, disaggregate activity-based models consider the more detailed movement of the logistic vehicles, but often disregard the commodities being carried. Although disaggregate models are much more accurate for predicting the influence of commercial vehicles on traffic patterns, they offer little help in evaluating the more aggregate economic impact challenges. In this paper we take a valuable step in bridging the gap between the two seemingly divergent schools of thought. Using recent agent-based developments in transport modelling, we demonstrate how different agents can be added to the transport model’s commuter population. Shipper agents are those wanting to convey goods (commodity perspective), and assign the shipments in a market-environment to logistic service providers and, ultimately, Carriers. The latter is injected into the agent-based model as individual commercial vehicles executing the pickups and deliveries that result from typical routeoptimisation initiatives within companies (activity-based perspective).