Addressing the challenge of providing a sustainable public transport service that optimises available resources, and simultaneously offers the best possible service to users is a complex issue. One innovative solution emerging on the streets of Barcelona is the On-Demand Bus Service, also known as Demand Responsive Transport (DRT). It involves figuring out ways and options to implement it, with efficient and satisfactory real-time results.
What is Demand Responsive Transport? DRT isn’t your typical bus route. Instead, it’s a flexible, user-driven approach that responds in real time to the travel requests submitted by passengers. Think of it as a hybrid between a traditional bus and a ride-sharing service. This system uses advanced algorithms to optimize routes based on where and when users need rides, striking a balance between minimizing travel time and meeting passenger demands.
This ambitious project is the result of a collaboration between the Knowledge Transfer Unit at the Centre de Recerca Matemàtica (CRM) and Transports de Barcelona SA (TB). The project team includes experts from both institutions, who have pooled their knowledge to test and refine this dynamic transport model.
The Mechanics of the On-Demand Bus
At its core, the On-Demand Bus operates based on requests that specify an origin and a destination. The intelligent system behind the service evaluates each request to determine if it can be accommodated efficiently without disrupting the service to other users. It’s a complex balance of optimization that requires innovative thinking to develop ad-hoc optimization algorithms.
To ensure the DRT system could deliver, researchers developed a simulation lab to test various scenarios. This simulation uses historical data to mimic potential passenger requests and evaluates the system’s responses under different conditions. The goal? To ensure reliability in terms of both passenger satisfaction and operational efficiency.
The primary way to validate the qualities of a service model is through its use by users. At the same time, this model is mainly described by a set of events or journeys that need to be met by certain users. We call them service requests, and the challenge of this stage has been to generate them so that they describe extreme scenarios that ensure user satisfaction in real time. Generally, the information available is given by the history of validations that have been carried out and registered with transport tickets. The descriptive analysis of the historical data provides a large volume of information, which has been considered. However, the key has been to generate a set of information that is unknown, such as the users’ destinations.
In a practical case of a conventional bus line, the study includes:
- Comprehensive analysis of the line’s schedules.
- Extraction of the temporal distributions of validations during a service day.
- Acquisition of the temporal distributions of daily boarding stops.
- Calculation of the duration distribution of a journey between any pair of stops on the line.
With this information, service requests are generated. At this point, each request must be assigned an entry time, an exit time, and carefully chosen origin and final stops. First, the entry time is determined from the temporal distributions of validations during the day; the times when more people use the service are more likely. Similarly, the origin stop is obtained from the validation distributions of the stops; the stops that receive more validations are more likely.
The effectiveness of the DRT model hinges on its ability to handle real-world complexities. The CRM team created scenarios based on extensive data analysis—from bus schedules and boarding patterns to journey durations between stops. This rigorous testing helps predict how the service will perform under various conditions, ensuring that the system is both robust and adaptable.
Designing routes for the On-Demand Bus is akin to solving a complex puzzle, similar to the “travelling salesman problem” but with added constraints like time windows and bus capacity. The system employs advanced algorithms, including A*, to find the most efficient paths through the city’s maze of streets, thereby ensuring that every route is as optimal as possible.
In the case of Barcelona’s transit system, researchers initially focused on 24 distinct bus lines. The challenge extended beyond simply analysing these routes individually; it also involved determining how they could be synergistically combined to enhance overall satisfaction and efficiency.
The inherent flexibility of the On-Demand Bus (DRT) model enabled consideration of various combinations of these lines. Given the vast number of potential combinations—literally thousands—the team employed specific filters to manage this complexity effectively, focusing their efforts on viable aggregates. This approach led to the creation of 272 distinct groupings, interpreted as individual lines or sets of lines, optimized to deliver the best possible outcomes in efficiency and user satisfaction.
Although this detailed analysis was specifically tailored for Barcelona, the methodology and insights gained have broader implications. The structured approach used in this study is adaptable, offering valuable strategies for other cities looking to optimize their transit systems in similar ways.
A Future Built on Flexibility
One of the most exciting aspects of the DRT system is its potential for scalability and adaptation. In Barcelona, the CRM has explored how different bus lines might be combined to enhance service efficiency and passenger satisfaction. This approach could easily be adapted to other cities, showcasing the versatile potential of on-demand transit solutions.
The On-Demand Bus project in Barcelona is more than just a transportation solution; it’s a glimpse into the future of urban mobility. By blending technology with traditional transit, cities can transform how we think about and use public transportation. As this project continues to evolve, it may well set a new standard for cities worldwide, making every journey quicker, easier, and a little greener.