ABSTRACT

Forecasting potential demand for a new mass transit service can be difficult. Typically during the planning process, a travel demand model is developed, employing several travel attributes and socio-economic information, to estimate modal shares by means of stated-preference travel survey. Although this disaggregate modeling procedures have been customarily applied to predict changes in travel behavior following the inauguration of new services, results are limited and usually inconsistent with the actual demand when such services are in operation. In this research, a case study of the Bangkok Mass Transit System (BTS), a first major rail-based transit service in downtown Bangkok, was conducted. A probabilistic choice model was developed in the present study, based on revealed preference travel survey. The model was evaluated and compared to those calibrated using stated-preference travel survey when the system was in the planning stage. Preliminary results of this research showed that BTS ridership forecast substantially exceeded the actual ridership, probably due to the stated-preference sampling bias. Careful comparisons between parameter estimates from the stated- and revealed-preference data will provide a better understanding of this disparity.

KEYWORDS: Stated Preference, Revealed Preference, Discrete Choice Analysis, Bangkok Mass Transit System

Back to List of Papers