BEHAVIOURAL DEMAND MODELING AND VALUATION OF TRAVEL TIME. AGGREGATION PROBLEMS

In the context of this workshop, the aggregation problem was primarily one of statistical inference for predictive purposes. Dissatisfaction with geographic aggregation (and aggregate analysis in general) has grown in part from the realization that the geographical zones upon which aggregate modeling is based are not homogeneous populations and, hence, are not appropriate units for aggregation. Disaggregate, stochastic transportation choice models have developed rapidly in recent years. These models estimate the probability that an individual traveler will select a mode, based on characteristics of the traveler, the environment, and the available choices. These disaggregate models have recently been applied to the prediction of future travel behavior. The basic problem of aggregation associated with use of disaggregate stochastic models appeared at this point. The problem was the use of such models to expand individual probability-of-choice estimates to describe travel-choice behavior of groups. Disaggregate modeling did not create the problem; it merely made it explicit. An additional issue discussed in the workshop was the role of attitudinal variables in explaining travel choice. Research is currently under way to investigate the potential of attitudinal variables for predicting travel behavior. In this research, there is a problem of separating groups of travelers according to their attitudes and other characteristics and of disaggregating so as to maximize between-group and minimize within-group variance.