Modelling bus bunching and holding control with vehicle overtaking and distributed passenger boarding behaviour

•We develop a bus motion model taking into account of bus overtaking, bus capacity constraint and dynamic redistribution of passenger queues.•We develop quasi first-depart-first-hold (FDFH) principles for headway- and schedule-based bus holding controls with overtaking to minimise the deviations from the scheduled headway.•We show that neglecting overtaking and passenger behaviour underestimates the benefits of holding control policies.•The benefit of overtaking is greater under higher stochastic travel time and smaller design headway, thus applicable in the high-frequency service.

Headway fluctuation and bus bunching are commonly observed in transit operations, while holding control is a proven strategy to reduce bus bunching and improve service reliability. A transit operator would benefit from an accurate forecast of bus propagation in order to effectively control the system. To this end, we propose an 'ad-hoc' bus propagation model taking into account vehicle overtaking and distributed passenger boarding (DPB) behaviour. The latter represents the dynamic passenger queue swapping among buses when bunching at bus stops occurs and where bus capacity constraints are explicitly considered. The enhanced bus propagation model is used to build the simulation environment where different holding control strategies are tested. A quasi first-depart-first-hold (FDFH) rule is applied to the design of schedule- and headway-based holding control allowing for overtaking, with the objective to minimise the deviation from the targeted headway. The effects of control strategies are tested in an idealised bus route under different operational setting and in a real bus route in Guangzhou. We show that when the combined overtaking and queue-swapping behaviour are considered, the control strategies can achieve better headway regularity, less waiting time and less on-board travel time than their respective versions without overtaking and DPB. The benefit is even greater when travel time variability is higher and headway is smaller, suggesting that the control strategies are preferably deployed in high-frequency service.