Schedule-based transit assignment model in transit networks with recurrent uncertainties

This paper proposes a novel within-day dynamic transit assignment model in networks with both demand and service uncertainties. The transit demand uncertainty is due to day-to-day demand variation and passengers' stochastic arrival patterns while the transit service uncertainty is result from the stochastic dwelling times at stops and transit vehicle travel times on roads. The schedule-based transit assignment approach is adapted for modeling passengers' risk-taking behavior on their route choice problem in dynamic transit networks with uncertainties. A dynamic network loading procedure is carried out in a time-incremental manner embedding a stochastic process. The proposed model is formulated as a fixed-point problem in a diachronic network. A multiple successive averaging algorithm is designed with propagation on passenger loads and service configuration. The proposed model is finally tested in a small-size artificial transit network for illustration.