By Bobby Du & Paul-Antonin Dublanche
An efficient and reliable public transit system plays an important role in mitigation of congestion and attraction of more users from private car. However, sophisticated traffic condition and dynamic travel demand often make public transit services unstable and uncertain, which results in longer waiting time especially during peak hours or special events. One common phenomenon called bus bunching (BB) or platooning usually happens when the headway between successive buses arriving at the same bus stop is less than the scheduled headway or a certain threshold. BB is a major source of congestion, which not only causes passengers’ travel time delayed and extra waiting time, but also degrades the bus operation performance. Most of the prior researches on BB were limited in a single or multiple bus lines, consequently, only a few studies were found that focused on the whole bus network in a city or even larger region. Recent advances in big data create new opportunities for exploring BB problem in a large-scale scope.
To tackle the BB issue, four questions need to be answered: When and where does BB happen? Why does BB happen? How to predict BB? What kind of strategies are able to reduce/avoid BB? As the first step, it is important to identify the bus stops and bus lines with serious BB issue as well as their features. Based on our preliminary findings from historical Opal card data analysis, the occurrence of BB shows its patterns at spatial-temporal-operational dimensions. At the spatial dimension, BB is usually identified at high density residential areas, school zones and central business district with high travel demand and multiple bus lines involved. At the temporal dimension, BB happens much more during peak hours than non-peak hours, more on weekdays rather than weekends. At the operational level, the frequency of BB is highly correlated to the features of bus stop, bus operator and bus route. For example, BB happens much more at stops with multiple bus lines (4 or more) rather than few bus lines (3 or less), and also more serious at interchange stops than non-interchange stops. Moreover, some bus operators keep buses running smoothly, however some bus operators face serious BB problem.
Although our research is based on historical data in 2016, which is kind of out of date and we expect the situation has been improved so far, it is meaningful and valuable to develop a series of methods to identify, analyse, predict and solve the BB issues. With the historical data, our methods have been tested and validated, and the findings are quite useful to narrow down our research focus to certain bus stops and bus lines to investigate the reasons behind BB. In the following work, the rest three questions will be answered by exploring the reasons behind BB, predicting the BB in real-time, and designing the corresponding solution strategies.
For more information visit the SMART Seminar event page.