In this paper, the problem of accurately describing patterns of intermodal competition in a situation in which there are a large number of alternative modes available is discussed. This research was motivated by effrts to increase the capacity and usage of the existing hudson river crossings connecting manhattan and northern new jersey. This corridor is characterized by the presence of an unusually large number of distinct transportation options and a high level of transit use. In such a setting, it is important to know not just how many commuters might use a new service, but also from which existingservices they would be drawn. The mathematical structure of an innovative model developed for nj transit and the port authority of new york and new jersey to allocate demand across seven primary modes ispresented. The representation of intermodal competition that this model provides is considered, and its properties are contrasted with those of some commonly used variants of the familiar logit model. Empirical estimates of the own- and cross-elasticities of demand implied by the model coefficients are broken down by mode, service attribute, and geographic area. This paper appeared in transportation research record no. 1139, Urban travel forecasting. For covering abstract see IRRD no 817822.
Abstract