Ride-Pooling as Component of an Urban Solution Strategy
‘Mobility as a service’ offers aim to persuade consumers to do without their own cars over the medium to long term. One feature is the use of pooling services, which combine similar passenger routes - with small detours - in order to offer a lower fare. This offers cities the potential to significantly increase the number of kilometers per passenger travelled on the roads while reducing the number of cars required. Ultimately, it uses the existing infrastructure in a more efficient manner. Since mid-April 2019, the city of Hamburg has been testing the effects of pooling services on the mobility usage of its residents in cooperation with MOIA, the new ride pooling service of the VW Group (Source:Hamburg.de). Pooling services can also be used as shuttle services for local public transport. The focus is on areas with inadequate connections to underground and suburban trains. The concept - long known as shared taxis in sparsely populated rural areas - is currently being tested by Deutsche Bahn and its subsidiary IOKI in Hamburg. If providers then also employ a fleet of electric cars, a positive effect for cities from an environmental perspective will also become apparent.
For cities and ride pooling providers, however, their success depends on the vehicle occupancy rate. If trips cannot be combined, there is no reduction in urban traffic, and it becomes increasingly difficult to cover operating costs. Therefore, it is necessary to identify passengers correctly for a joint journey and, from the consumer's perspective, to transport them without loss of comfort, which is measured here in terms of the time required for a route. This decision is made using an algorithm that takes a variety of factors into account. Depending on the departure and arrival points, the traffic situation, but also the weather, trips are ultimately pooled in the best feasible way. Furthermore, the supply of trips has to be optimally balanced with current demand. This is because every pooling car that spends too much time on the road reduces the overall load factor. The trend here is towards self-developed, complex forecasting models. Nowadays, the computing capacities at our disposal make it possible to implement self-learning forecast models based on machine learning – and mostly cloud-based - approaches.