There is a wide perception that freight models need a careful review, updating and overhaul to meet current requirements. This is partly true, but the context of the applications has a major influence, and in many areas the current methods are adequate for at least some purposes. Freight and commodity flow models are used in a wide range of situations. Some of these are best described as 'predicting the present'. Freight vehicle flow matrix estimation, as it is frequently used, is a good example of this, and is entirely satisfactory for many user requirements. Dissatisfaction arises when the match between the expectations of the end users are not well matched to the constraints inherent in the model. There are several quite different types of freight model, each suited to a different group of applications, and none of which are suitable for all. Reviews over recent years have bee oriented towards national and urban transport planning, and so adopted a comparatively narrow view of 'freight models'. range is rather wider. Traffic planning and management have substantial requirements for freight vehicle and movement modelling. Logistics operations and optimisations provide a very different and supply driven view of movement patterns, and capacity management requires detailed simulation of the wide variety of vehicle movement and response characteristics. There are many other specialised areas, from shipment choicer and empty running estimation to the unresolved role of service industries on urban freight movements. Models aimed at each of these areas have been reasonably effective to at least some extent in their limited domains, although there are clear shortfalls in modeling capacity, data supply and validation even then. The unease about shortfalls in freight modeling is focussed more on issues of forecasting and demand estimation than the operational areas. The distinctions between four major categories are used to identify where efforts may best be applied to improve the current state of practice, and of theory in modeling. Both are critical to transport planning and management. However the major distinctions are drawn between the following categories: Predictions of the present Models aimed at pivot point analysis of variations in freight conditions or demand. Models aimed at predictions of the present and recent past. Forecasting models. These distinctions are very helpful in identifying where improvements in practice or theory are needed, but the management demands of the road and transport system now demand real time and monitoring capacities as a major function. These demands cross over the above categories, and make new demands. The decision making elements in overall transport systems are no longer always optimal, or even rational, unless a satisficing approach is taken. If the emphasis is moving towards monitoring and adaptation then the types of models required will be more and more data intensive and be linked to optimal control approaches, not yet common in the transport arena. This and other modeling approaches demand data from ITS sources, and make good use of them. At the other end of the scale, it is not necessarily possible or indeed appropriate to exclude decision making parties - who may not be entirely rational or working on full information. These areas may therefore require models including elements of agent behaviour. The paper considers each of these areas progressively, and drawn conclusions on where data, now theories and different approaches to modeling and problem formulation may be the most appropriate to meet new and emergent demands in each domain. For the covering abstract please see ITRD E135207.
Samenvatting