Georgia Tech researchers address bus bunching

Georgia tTech researchers address bus bunching
Bus Bunching Demo

As any city dweller knows, buses are rarely on time. It’s typical to wait a while, only to have several buses show up one after another, a phenomenon known as bus bunching.

Fortunately, researchers and students at the Georgia Institute of Technology have developed a possible solution for bunching that provides better service to riders, simplifies the job of drivers and reduces work for management.  

Industrial & Systems Engineering Professor John Bartholdi and Georgia Tech alumnus Don Eisenstein (MS IE 1983, PhD IE 1992), a colleague from the University of Chicago, have devised a way of computing the delays so gaps between buses “self-equalize.”

Now an interdisciplinary team of Georgia Tech students have built a system of tablet computers to control the trolleys using the equation on Tech’s main campus route, which carries more than 5,000 passengers a day.

“Because of its simplicity, our scheme is easy to implement and easy to adapt,” said Bartholdi, who is also the Manhattan Associates Chair of Supply Chain Management. “We expect it to be useful for other transportation systems with short headways, such as subway trains or airport shuttles.”

The first step toward reliable bus service, Bartholdi says, is to abandon the fixed schedule and have drivers go with the flow of traffic.

Under the “self-equalizing” plan, each bus is equipped with a GPS and cellphone. The GPS constantly reports the bus’s position to a central server. When the server recognizes the bus has reached a stop, it sends a message via cellphone telling the driver how long to wait and when to proceed.

That departure time is calculated through the “self-equalizing” equation, which changes the headway of each newly arrived bus to an average of its former headway and the headway of the trailing bus. For example, if its former headway was larger, its new headway becomes smaller. Using the equation, gaps between buses will equalize even if a bus is added or removed, or if the bus route changes.

“The equation computes a wait time for each bus arriving at a control point in such a way that gaps between buses tend to equalize,” Bartholdi said. “Exactly how that happens is the magic of the mathematics.”

When the research team tested the system on Tech’s bus line earlier this semester, the researchers received positive results and favorable reviews from bus riders and drivers.

This is a Vertically Integrated Project (VIP) at Georgia Tech, which combines students from across disciplines to solve a real-world project. The research outcomes will be published in Transportation Research Part B in May.

Explore further

Using math to speed up school buses

More information: Read more about Self-Coordinated Buses Resist Bunching here .
Citation: Georgia Tech researchers address bus bunching (2012, April 19) retrieved 17 September 2019 from
This document is subject to copyright. Apart from any fair dealing for the purpose of private study or research, no part may be reproduced without the written permission. The content is provided for information purposes only.

Feedback to editors

User comments

Apr 20, 2012
Great... so instead of just going when there aren't any more people waiting, you'll sit at the stop anyways.... making everyone on the bus late.

Great idea. Not.

Apr 23, 2012
Drum-Buffer-Rope. Old concept, good new application.

Apr 23, 2012
Great... so instead of just going when there aren't any more people waiting, you'll sit at the stop anyways.... making everyone on the bus late.

Great idea. Not.

Perhaps I am missing something, but the term "equalize" suggests that wait times will self-adjust to the optimum value. The outliers are removed from this system -- the whole idea is to prevent one bus from being late and the other early. Not only that, but the control points are at the end of the line, where there are generally no passengers, or at transfer stations, where additional wait time provides an advantage. Indeed, the results from real-world testing were reduced wait times overall.

This is a great idea, particularly if it can be paired with more intelligent traffic management systems.

Apr 23, 2012
City buses stink, they block vision and they slow all traffic down. They're "special" and get to stop in no-stopping zones, backing up cars in that lane, causing them to merge left, which slows that lane down too, which slows the whole stream down. They also cause more traffic to have to stop at more red lights, causing more cars to idle unnecessarily. They also don't pay for themselves, causing a drain on the public coffers.

The best normal traffic days in recent memory were when the bus drivers went on strike. Traffic flowed wonderfully.

City buses suck.

Apr 23, 2012
Strongly recommend Gerald Mars' book "Cheats at Work: An Anthropology of Workplace Crime". There is a cultural anthropological reason for bus bunching. The issue isn't a mathematical algorithm. It's the drivers.

Ever wonder why it seems to take 15 minutes to catch your waiter's eye to get the check at a restaurant?

Look up donkey jobs.

Apr 23, 2012
Lord Jag, You are not taking into account how long you will have to wait for the bus in the first place. If their is bunching you wait for the bus to arrive will be far longer than your wait would be at each bus stop after you are on the bus.

Apr 23, 2012
"For high frequency bus runs where there is a bus every 10 to 15 minutes, fixed schedules are not needed."

Not true. When bunching occurs the wait time for the next bus rises above the 10 to 15 minutes, with most of the buses running empty that arrive immediately after another bus.

Please sign in to add a comment. Registration is free, and takes less than a minute. Read more