Segmentation of the transportation market with regard to activity-based freight transport modelling

Abstract

Transport models are used to estimate the expected effects of policy measures—such as infrastructure building, upgrading or the introduction of a highway toll—on the users of a transport system. The results of transport models are a necessary input for cost–benefit analyses, which answer the question of whether a policy measure improves the economic welfare. In passenger transport modelling, the activity-based approach links peoples' daily behaviour patterns to their movements through time and space. Traffic—as a macroscopic phenomenon—can be explained as the sum of many individual behaviour decisions. In the modelling of freight transport, the state-of-the-art method is still the aggregate and functional approach, which concentrates on regions/traffic cells and their gross domestic product by economic sector instead of individual logistical decisions. The transports are losing the relation to the responsible decision-maker. There is still a lack of suitable actor-based micro-models that are capable of adequately modeling freight transport systems. Having a microscopic modelling base for the transport of goods would be a significant improvement for transport forecasts and the assessment of policy measures, because of its ability to map individual adaptation reactions. The article gives a short introduction to microscopic modelling techniques in freight transport. However, the question of the modelling object and its classification is the first to be answered when modelling and simulating shipper and carrier behaviour. Relevant data sources are critically analysed with regard to activity-based freight transport modelling. It is shown that inadequate classifications have detained methodical progress in freight transport modelling. A suitable classification schema is deduced, which is basing on the notion of transport markets. For the identification of the characteristics of these markets from existing trip samples, fuzzy-clustering techniques are presented. Further approaches for calibrating a first generation of activity-based freight transport models using the existing data sources are shown.