Many city-dwellers are familiar with what researchers call "bus bunching." It's the "Wait all day for a bus, then two come at once " phenomenon. Two engineers say the way to fix the problem and reduce wait times overall is to abandon bus schedules altogether, where buses arrive frequently. "No one cares about a schedule as long as the gaps between buses — that is, the headways — are small, as in a busy urban bus system," the researchers, John Bartholdi III of Georgia Tech and Donald Eisenstein of the University of Chicago, wrote on their project's website.
Instead, they think, buses should check into headquarters constantly and strive to maintain equal distances between one another by pausing as needed at bus stops. The engineers tested their idea in a 3.3-mile Georgia Tech campus shuttle route and published their results in the May issue of the journal Transportation Research Part B.
Bus bunching happens because especially long or short distances between two buses tend to get exaggerated, the researchers explained in their paper. A bus with a long headway in front of it will go slower and slower because the long gap brings more passengers to bus stops, slowing boarding. A bus with a short headway will move faster, making its gap even smaller, because it carries fewer passengers. With a major disruption, such as bus breakdown, it's difficult for behind-schedule buses to ever catch up.
To combat bus bunching, researchers installed Android tablets with a bus-control app in Georgia Tech campus buses that serve about 5,000 passengers a day. (In contrast, Chicago's buses serve about 1 million passengers a day and New York City's, more than 2 million every weekday).
The app constantly reported each bus' position, headway and velocity to a central server. At two "control point" bus stops along the route, the app sent drivers a text message telling them to pause, if needed, to maintain equal headways between all buses. Otherwise, researchers told bus drivers to go with the flow of traffic, instead of driving quickly or slowly to try to meet a schedule.
There was no target headway time. Instead, the system calculated headway times based on what a bus' previous headway was and the headway of the bus behind it. Times varied through the day.
Researchers found the new method reduced the length and variability of passengers' wait times. Bus drivers found it less stressful than adhering to a schedule, according to the project's website.
Researchers then took one bus out of the route during summer semester, to see how the system responded to disruptions. The remaining two buses equalized their headways and didn't bunch, they found.
This isn't the first time someone has suggested running buses based on their headways, rather than a schedule, but it is the first time anyone has tested the idea in real buses, the researchers wrote on their site. The idea should be easy for others to try, they said. "Our scheme is easy to implement and easy to adapt," Bartholdi said.
The scheme would work best in cities where buses run at least every 10 or 12 minutes, Bartholdi and Eisenstein wrote in their paper. A Chicago Transit Authority analyst told the engineers that passengers tend to ignore schedules, anyway, if the gaps between buses are that short.
Though the managers of major bus systems are "understandably reluctant to 'play,'" as Bartholdi and Eisenstein wrote, they suggested their system would work well for airport shuttles and the temporary bus routes cities set up for fairs and events.