No information can be more important to transport planners, designers, and analysts than reliable forecasts of the peak and off-peak travel volumes on transport networks. Yet, no reasonably complete and valid methodology has been proposed-much less developed and verified-that will permit the transport planner and analyst to properly differentiate travel by time of day; to determine realistically the duration and level of peaking and recognize its dependence on the transport system design and performance; and to account for shifts in trip-making from one hour to another, from one mode to another, and from car pooling to driving alone in response to changes in transport system design, in service, or in price. Further, the usual travel forecasting process treats trip generation as though trip- making were independent of transport system changes, and it treats the trip distribution, modal-split, and route assignment phases as though trip-making choices-whether to travel, final destination, mode, and route-are made sequentially and apart from the circumstances attendant with the other phases. Accordingly, the purpose of this paper is to formulate a model structure such that these phases can be treated simultaneously, that the total amount of trip making (as well as the destination, modal, and route choices) can be varied with the transport system and its performance and price characteristics (among other factors), that shifts from car pooling to driving alone can be represented, that shifts from one hour of travel to another can be characterized, and that the amount of travel during peak and off-peak hours (i.e., the absolute build-up or decrease in peak or off-peak flow) can be determined. /author/
Abstract