Impact of Service Design on Urban Ridepooling Systems

Ridepooling services promise to provide fast and comfortable transportation at an affordable price level. Several studies and simulations have shown the potential for positive impacts on congestion, pollution and consumed resources by replacing current car trips or taxi rides with a fleet of (autonomous) ridepooling vehicles. In this work, demand and supply data from a ridepooling company in Hamburg, Germany, are taken to build up a simulation in the framework MATSim. Approximately 14,000 requests occured on a particular day and were partially served by more than 100 vehicles in a 200km
²
service area. The impacts of pooling parameters (e.g. maximum wait time or allowed detouring), scaled demand and supply and stop network density are assessed in regards to served requests, produced mileage and service quality. The simulation shows that the used pooling parameters have a high impact on the system efficiency. The adjustment of the maximum waiting time or the maximum allowed detour can increase or decrease the pooling efficiency by around 20 %. For the stop network assessment, the given ridepooling stop network is compared with a door-to-door service and the current public transport stops. The results demonstrate that a thinner stop network leads to less empty mileage and more rides served, but customers face longer walking.