Modeling Bus Service Time for a Curbside Stop

Abstract

Bus queuing occurs frequently at the entry and exit areas of curbside stops. Formed queues can induce extra delays for bus operations. Conventional regression-based models to analyze bus service time cannot capture such delays because of their limitations in addressing interactions between buses and arriving passengers. To capture the extra delays, this paper proposes a new approach to estimate bus service time on the basis of the Monte Carlo method. The proposed models account for interactions between arriving buses as well as the numbers of boarding and alighting passengers. The models were established for curbside stops with both one and two berths. Case studies were implemented to show the effectiveness of the proposed approach. Archived data from the automatic vehicle location system and the automatic fare collection system were used to calibrate and validate the models. With the established models, the impact of passenger arrivals on bus service time was further demonstrated.