On-demand transport: one or hundreds?

Fortunately, on-demand transport has become a widely known concept with both consumers and the transportation industry. Whether an on-demand service is fully integrated with our cities’ and regions’ transit systems, or whether they are independent taxi or van-based services, on-demand operations are increasing more and more, especially since mobile-based apps allow operators to gather users’ preferences, calculate the desired routes and send drivers to clients in real-time.

There is no single kind of on-demand transport, but a wide variety of models. On-demand transportation can be defined as any type of transportation service whose offer (schedule and/or route followed) is not fully defined until the operator receives some sort of information about the user’s intention, for that specific day, place and/or hour.

Under this definition, at Shotl we like to describe 5 groups, or levels, of on-demand transportation:

1. Fixed route: At this level, on-demand transportation services operate with a predefined route, where some stops are defined as “optional,” so the vehicle only travels there if somebody has requested that specific stop earlier. This allows the operator to save on distance travelled empty and on travel time. Sometimes, the whole route can be defined as optional, so if there is no demand for a certain trip, the vehicle doesn’t even leave the garage.

2. Flexible route: The previous case can be expanded by adding optional stops in a buffer area near the route followed. If a stop is requested, the vehicle detours. Each trip has a maximum detour it can reasonably assume, so if this maximum has been reached, users can choose to wait for the next bus or to walk to the “main route” stop.

3. Feeder: Here we add an extra level of freedom, where all the pickups (or, alternatively, all the dropoffs) are defined based on the demand, not following any set route. However, the final point (or, alternatively, the initial point) is set at a certain place and time. After this time, the trip can start again.

4. Corner to corner: In such a model, we completely remove the concept of a trip. There are no predefined routes nor schedules, but a set of stops distributed sparsely in an area and a time window where there is service. The route is created in real-time based on the users’ petitions. If there is no demand, the vehicle stops. If there is constant demand, the vehicle can keep moving until the end of the service time.

5. Door to door: The case above still requires the definition of stops, either physically or virtually, as the only allowed points of pickup and dropoff. The last step is to totally remove this concept, and allow pickups and dropoffs everywhere within a service area.

Use cases for each of these models are very diverse: while “corner to corner” and “door to door”, the specialities of Shotl, are ideal for suburban areas, night services, company campuses and small rural and touristic areas, the “feeder” model is ideal for an airport shuttle or a company bus, and fixed & flexible routes work fine in large, busy areas. Also, mixed models can easily coexist.

This classification ranges from the least to the most flexible models of on-demand transport, but this is not the only criteria to classify them. Another important model looks at how much in advance the route is calculated, estimated times are provided to the customer and new petitions are scheduled. This depends on the unpredictability of the demand: a bus to school will only change every new school period, a booking for a shuttle to the airport can be done a few hours in advance, and users of a night service will highly value last-minute booking.

Algorithms that are constantly re-optimizing routes are ideal for the last example, where algorithms that find the optimal solution only when all the demand is known will be better for the former.

Finally, there are external factors that might define on-demand: the type of vehicle (a bus, a van or a taxi), its integration (or lack of integration) with public transport fares and information, the nature of the destination, and so on. If you’re thinking about implementing an on-demand transport system, contact us and we’ll help you decide the ideal mode for you.

Popular posts

Read more

28.06.21

Gender equality and sustainable transit are connected – here’s why

Sustainable transit and gender equality are central to many of the UN Sustainable Development Goals (SDGs). Like the proverbial horse and carriage, though, you can’t have one without the other.


Adrià Ramírez
Read more

29.03.22

3 regions leading the way with on-demand transit

At Shotl, our experience shows demand-responsive transit (DRT) works best in suburban or rural areas that are underserved by traditional fixed-route/schedule transit. With this in mind, we take a look at three European regions that are leading the way


Adrià Ramírez
Read more

30.09.19

Machine learning boosts Shotl users’ experience

Machine Learning (ML) is a field of Artificial Intelligence (AI) which focuses on the development of techniques that allow computers to learn how to perform a specific task.


Enric Sanmartí
;
Subscribe to our Newsletter