Chapter 11: The Estimation of Origin-Destination Matrices in Transit Networks

The steady-state transit assignment problem is usually modeled by a headway-based approach, in which the arrivals of transit vehicles on a transit line are described by means of its average headway. Le Clerq (1972) was among the first to point out the problem of common lines on transit networks. A large body of literature has since been devoted to solving this problem (Chriqui and Robillard, 1975; Last and Leak, 1976; Nguyen and Pallottino, 1988; De Cea et al., 1988; De Cea and Fernandez, 1989, 1993; Spiess, 1987; Wu et al., 1994). However, little effort has been spent on dynamic trip assignment for transit networks. The headway-based approach assumes that each transit line operates at a constant headway, and that the speed of a vehicle is determined by a volume/delay function. The passenger waiting time for boarding a line is a probabilistic function of the vehicle headway. This is a static model that assumes a steady state situation and predicts the average performance of the transit system within a specified period, e.g. at peak hour or off peak hour.