A Logistic Approach for Freight Modal Choice Model
NUZZOLO A, University of Rome, RUSSO F, University of Reggio Calabria, Italy
In recent years, numerous theoretical works have been produced which tend to unify the information criteria of freight demand levels, passing from completely descriptive model systems to miÃ?ed model systems which entail the use, for certain choice dimens
In recent years, numerous theoretical works have been produced which tend to unify the information criteria of freight demand levels, passing from completely descriptive model systems to miÃ?ed model systems which entail the use, for certain choice dimensions, of behavioural models.
The purpose of this paper is to report, a behavioural freight modal choice mode], obtained within the framework of a research aiming to construct a system of freight demsnd models on a national scale (Cascetta et alii 1995; Nuzzolo et atii, 1995). The project is being financed by the Italian National Research Council (CNR) in the 2nd Finalised Transportation Research ProgrRm (PFI'2).
In the Italian project a modelling architecture combining Spatial Input-Output models (Moses I955; Chnery and Clark, 1959; Costa, 1973) with random utility mode and path choice has been adopted for reasons of flexibility, data availability and internal consistency.
The Multi Regional Input-Output model produces region-to-region trade flow matrices for each economic, sector using final demand vectors, technical coefficient m~trices for regional economies and inter-regional trade coefficients which are elastic with respect to the generalised transport costs. (Cascetta et alii, 1996).
These matrices expressed in monetary terms are transformed into Origin-Destination matrices Of freight quantities through specified indices of transformation/amplification. (Rnsso and Conigliaro, 1997),
In the subsequent step the system contains an aggregate descriptive model that splits Origin- Destination matrices of freight quantities; with regard to dimensions, through exogenous coefficients into "captive" large shipments via sea or block train, and unconstrained small and medium size shipments.
The O/D matrices of small and medium size shipments were split among subregional zones (provinces) belonging to each region following a gravity model.
Then a macroclass identification model allows aggregation of the quantities of goods belonging to the different sectors obtained by Input/Output models. It is supposed that there are behavioural homogeneities in decisions regarding freight shipments belonging to the same macroclass of goods. Three microclasses are currently used with regard to freight: perishables, consumer goods and capital goods.
In the current model system a descriptive size classification model allows us to obtain, for each freight macroclass, the percentage distribution of the whole shipment into a finite number of weight classes. A firm's modal choice model, simulates the distribution among modes of road transport, rail transport, combined transport and shippers. The recent work have been developed to perform this last step in a different way. This paper describes the firm's modal choice model in a logistic context, obtaining within the same model, size and mode chosen.
The next section reports a general theoretical framework of the modal choice problem and the main approaches are outlined. Then the structure of the logistic model for the simulation of the size and mode chosen by the firm will be described. The model will be dealt with through a disaggregated approach which explicitly considers as decision makers the firms, with a random utility model of choice. Finally, the results of specification and calibration of firms' logistic behavioural choice models, obtained by using the data from sampling surveys carded out in Italy, are reported.
Association for European Transport