A group of engineering students and professors have developed an algorithm that can predict the pace and timing of bike borrowing and returns in Tel Aviv.
By Lital Levin

While visiting Paris a few years ago, Iris Forma rented a bicycle from the Parisian municipality, run under the same system that will soon launch in Tel Aviv. When she arrived at her destination and wanted to return the bike to the nearest city bike rack, she discovered that it was full and no lock was available.

The municipality did not necessarily expect that there would be a natural balance between the number of bikes and destinations; like all other cities that have set up community bicycle systems, Paris operates a fleet of vehicles whose sole purpose is to shuttle bikes from one destination to another, to match supply with demand. Still, Forma was forced to wait a long time until a space in the rack became free.

“It is an excellent system,” she said this week, “but when there are breakdowns like this in service, it loses its effectiveness.”

And if such hitches occur in Paris, what can we expect in Tel Aviv? Forma, 33, a doctoral candidate in engineering at Tel Aviv University, believes there is a solution. She and her thesis supervisors, Prof. Michal Tzur and Dr. Tal Raviv, along with masters degree candidate Dana Pesach and other students in the department, have developed an algorithm that can predict the pace and timing of bike borrowing and returns at various stations across the city.

“Naturally, there will be an imbalance between the number of bikes at different stations,” Pesach and Forma explain. “In other networks around the world, the problem is solved through guesswork, or by putting out fires – when one station is filled to capacity, cars in the fleet are called in to remove the bikes, or the opposite.”
Traffic considerations

The mathematical formula developed by the Tel Aviv University team takes many variables into account: a station located at the top of a hill, for example, will see fewer returns depending on the time of day; in the morning, there will be a greater demand for empty spaces on racks close to universities, so that students can park the bikes on which they rode to school; by contrast, in the afternoon, when the students leave the university, there will be a greater demand for bicycles.

To avoid an absurd situation, in which the vehicles transporting bicycles cause traffic jams on their way from station to station, and when they stop to load or unload bikes, the model also takes into account traffic on the roads at certain hours of the day.

“It’s possible to transport the bicycles both during the night and during the day,” Forma says. “These are two different approaches and our study incorporates both of them. The advantage of moving the bikes at night is that traffic flows freely in the city at those hours, and the system’s vehicles can therefore move quickly without impeding traffic. On the other hand, transport by day allows a more efficient response to urgent situations.”

Pesach has been working in recent months with the FSM company, which set up the bicycle lending system in Tel Aviv.

“We are happy with the collaboration,” Pesach says, “and there is a great deal of interest in our model.”

The rest of the world is showing interest, too. The Tel Aviv University team recently received statistics from Washington, D.C., in order to develop a model for the shared bicycle system in the U.S. capital.