STRUCTURAL TRAVEL DEMAND MODELS: AN INTERCITY APPLICATION

Conventional sequential transportation models clearly have limitations as estimators of intercity travel demand. Despite their theoretical advantage, little work has been carried out in the full application of behavioral or models. Structural-model development is focused primarily on disaggregate models, particularly for modal split. This paper discussess the development of an alternative approach, that of developing a set of direct-demand models for estimating intercity transit travel for a Sacramento-Stockton-San Francisco Bay Area corridor study. A series of judgments are desired that identify why structural models rather than sequential models were chosen and why direct-demand models rather than probabalistic-choice models were used. The methodology of calibration, including variable selection and equation development, validation, and forecasting, is outlined. Emphasis is placed on the trade-offs to be made among policy responsiveness, accuracy, and the practical problems of developing and using such forecasting tools. The material has been oriented toward the planner-engineer faced with the practical issues of selecting and using intercity travel demand forecasting procedures.