How to make an on-demand bus network viable

MASTER'S PROJECT - For her EPFL Master’s project in civil engineering, Gaelle Abi Younes examined why on-demand buses have met with only limited success. She used technical and financial models to show how they could become a genuine alternative to cars in outlying areas.

Thanks to the ubiquity of smartphones, we now have a wide range of on-demand services to choose from for traveling short distances: bike-sharing programs, ridesharing, Uber, Lyft and more. But these services are generally centered in urban areas and less available in outlying areas or along main railway lines. The only option for serving people in these areas – and giving them an alternative to personal cars – is to establish a system of on-demand buses. Such systems, few in number, are what Gaelle Abi Younes decided to focus on for her Master’s project.

“I wanted to study a topic that was modern and innovative and that could make a real impact,” she says. “On-demand buses checked all those boxes. These buses aren’t new, but they’ve fared poorly in the past because not enough research and modeling had been done on them. There was a big gap in the literature, and that prompted me to look closely at this topic.” As an example, Gaelle Abi Younes points to various pilot projects that were carried out recently in the Lake Geneva area but weren’t taken further.

On-demand buses compete directly with personal cars rather than with conventional public buses.

Gaelle Abi Younes, EPFL Civil Engineer

On-demand buses are positioned somewhere between regular large-capacity public buses and personalized services like Uber. Each one can carry 25 to 30 passengers, and their routes are designed to link outlying areas to a public transport network or the closest urban center. “On-demand buses compete directly with personal cars rather than with conventional public buses,” says Abi Younes. “They offer a promising new way for countries like Switzerland to reach net-zero targets for transportation.”

Technical and financial modeling

In the first part of her Master’s project, she examined existing on-demand bus systems such as Ebuxi, introduced in Bern four years ago. She also reviewed available studies to get a better idea of the social and economic factors associated with on-demand buses, such as the waiting times that passengers generally deem acceptable and the type of people who might be interested in this type of service. “But what was missing was a comprehensive analysis of how to strike the right balance between cost, waiting time and transport time – even though similar optimization models exist and are widely used by companies like Uber,” says Abi Younes.

She therefore developed a model to simulate on-demand bus systems in two different countries – China and Greece – in order to see how the systems could work in cities of different sizes. The simulations let the civil engineer test the equations underlying her hypothesis and determine the conditions required to balance supply and demand. She used those results to describe how the systems could operate and to quantify how profitable and efficient they could be. “I found that the more people who use these systems, the better they’ll work,” she says.

With my models, planners could come up with a workable simulation, decide how to best invest their money and predict whether their system would work in practice.

Gaelle Abi Younes, EPFL Civil Engineer

For the second part of her project, she modeled the financial aspects of on-demand buses based on the same equations. This model was used to examine several scenarios, ranging from how bus operators could maximize their profits to, more realistically, how governments could subsidize on-demand bus systems to achieve passenger prices and travel times that are competitive with personal vehicles. “Today it would be very expensive to run a pilot project,” says Gaelle Abi Younes. “But with my models, planners could come up with a workable simulation, decide how to best invest their money and predict whether their system would work in practice.”

At the crossroads of several disciplines

Abi Younes did her Master’s degree in Switzerland after getting her Bachelor’s in civil engineering in Lebanon. She now works for TPF, the public transport operator in Fribourg Canton. “Switzerland is a benchmark in public transportation, with many different modes of transport used,” she says. “My Master’s project was even better than I expected because it gave me exposure to a variety of fields, from traffic-light modeling to transportation economics and sociology.” She hopes to have her project findings published in a scientific journal and may soon be able to apply them directly at TPF. The company has long been interested in on-demand buses and tested one last summer at Estavayer-le-Lac.


Gaelle Abi Younes, “On the modelling and analysis of on-demand micro-transit services,” Urban Transport Systems Laboratory (LUTS), EPFL. Master’s project supervised by PhD student Lynn Fayed, under the direction of Nikolas Geroliminis, EPFL full professor and head of LUTS, 2023.