Models for bus queueing at busy stops and corridors, and congestion mitigation strategies

Abstract

Bus agencies often operate bus lines with high frequency to meet the ever-increasing demand. Problems that come along with high bus flow and patron demand include long bus queues and poor service reliability. This thesis develops analytical and simulation models for estimating capacities and key performance metrics (i.e., bus delays and headway variations) for congested bus stops and corridors. At the stop level, built upon the previous analytical models for estimating isolated stops' bus-carrying capacities, this thesis develops analytical approximations to estimate the bus-carrying capacities at near- and far-side stops with one or multiple curbside berths. The approximations are derived using time-space diagrams of bus trajectories and probabilistic methods. They correctly account for the effects of key operating factors that were ignored or incorrectly addressed by previous methods. These factors include the signal timing and the distance between stop and signal. Comparison against computer simulation shows that our models furnish much more accurate estimates for near- and far-side stop capacities than previous methods in the literature. Numerical case studies are performed to examine how the stop capacity is affected by various operating factors. New findings and their practical implications are discussed. Looking at the corridor level, buses form queues at stops along corridors when the number of berths in each stop is insufficient to serve its bus flows. Buses might discharge from these queued stops at headways that are smaller than scheduled. This tends to induce longer queues at downstream stops and higher variations in the bus headways entering them. A vicious cycle of growing bus delay thus propagates along the corridor. Patrons suffer as a result. A simulation model is therefore developed in-house to jointly examine bus queueing dynamics and headway variability propagations in busy bus corridors shared by multiple bus lines. The model is used to examine two control strategies aimed at alleviating the vicious cycle in busy corridors. Both strategies hold buses at a corridor's upstream end. The first strategy does so to form bus convoys that are then released to traverse the corridor in unison. Though previously reported to be beneficial, present findings indicate that convoying generates greater headway variations and longer bus queues at stops than what occur under a do-nothing alternative. The second holding strategy releases buses into a corridor at fixed headways. This strategy performed surprisingly well, not only in regularizing headways, but in reducing bus delays. The strategy was found to be especially beneficial for long corridors with many queued stops and high patron demands for travel. Practical implications of the findings are discussed. Discussion includes means of harnessing emerging technologies to enhance the headway-regularization strategy.