1988 年 19 巻 p. 147-165
Since middle 1970's, many studies have been carried out on disaggregate travel demand models. However, applications of those models to transportation planning are limited: they are usually applied to the local and short-term transportation planning, or they are used in the frame of traditional 4-step travel demand forecasting models by replacing some of those steps. That is, there have been few attempts to develop and apply disaggregate travel demand model systems for predicting travel demand in metropolitan areas.
This paper develops a disaggregate model system for the travel demand of workers during a day in a metropolitan area.
Daily travel pattern by an individual consists of visits to one or several destinations in trip chainnings that are movements away from his base (home or work place). We assume that these decisions during a day are not independent each other but there exist some relationships among trips that make up tours. Therefore, it is important to represent these relationships among trips in modeling the daily travel demand.
In this study, by using the concept of utility maximization of base-to-base trip chainnig, we attempt to develop a set of models for predicting trip generation type, trip purpose, destination, mode choices in each trip on the way from work to home. The structure of travel demand model is represented by using nested logit model so that we avoid a simultaneous model's complexity caused by a large number of alternatives that a traveler faces in making his travel decision, while the model represents simultaneous decision-making process. The model incorporates the concept that decisions of a trip in a trip chain depend on decisions of trips conducted theretofore and decisions of trips planned thereafter, as well as on current trip conditions.
The estimation of the model parameters was carried out by using a sequential approach. Empirical examinations of the model system for Nagoya metropolitan area show encouraging results.