Demand Forecasting Models for Urban Goods Movements

In order to reduce the impacts of freight transport, Public Administration usually implements city logistics scenarios, whose effects have to be evaluated by using ex-ante assessment procedures. This paper proposes the state-of-the-art of models for the prediction of urban goods movements, which have been developed to support the above ex-ante assessment. Aiming at preserving the economic sustainability of the businesses located in the city and, at the same time, the environmental quality, the new challenge of urban traffic management is the optimization of the access of a large number of trucks and vans delivering goods in the urban area. Thus, the availability of a reliable tool for ex-ante assessment plays a key role in the decision making processes. Starting from the structure of urban goods distribution and its critical states, this paper identifies the decision-makers, whose choices could be influenced by city logistics measures. Furthermore, considering the outcomes and goals to be reached by Public Administration, the city logistics measures are also classified. In such way, it is possible to define the decisional processes influenced by city logistics measures, that models have to simulate. The presented modelling system allows us to forecast the OD truck flows within the study area and it consists of two sub-systems: the first related to the demand and the second related to the logistics. The former gives the OD matrices in terms of deliveries; the latter allows to convert the delivery OD flows into truck OD flows. The demand sub-system has been specified as a partial share model. From socio-economic data of the study area, it allows us to estimate the OD matrices in quantity characterized by service type, as well as the OD matrices in deliveries characterized by time slice and vehicle type. The logistic sub-system allows to estimate the OD matrices in vehicles. The modelling framework consists of two models that allow us to reproduce the commercial vehicle tours within the urban area. The former model gives the distribution of tours per number of deliveries; the latter gives the probability to choose the following destination for the next delivery.