Service Efficiency of Ridepooling Systems for Different Relations of Demand and Fleet Size

The main objective of on-demand ridepooling is to merge multiple requests for a single vehicle. However, sufficient demand for the service is crucial to achieve this goal. Hence, the service’s efficiency highly depends on the number of requests. On the other hand, providers need to decide how many vehicles to use for their service in order to achieve the highest possible occupancy rate while keeping the percentage of empty, idle vehicles low. Therefore, the knowledge about how many vehicles are required to serve a given demand is of particular importance. In addition, whether a given ratio between the number of operating vehicles and the number of requests results in satisfactory service efficiency is of great interest, especially to providers. Addressing these issues, we present a simulation study of an on-demand and stop-based ridepooling system, evaluating 1000 different combinations of fleet size and demand during the peak hour in a mid-sized German city, considering realistic traffic demand. In our evaluation, we focus on crucial efficiency criteria such as waiting time, occupancy rate, number of rejections, and share of empty vehicle kilometers. Our results show a clear relation between demand and fleet size for the efficiency of a ridepooling service. In particular, we found that a constant ratio between demand and fleet size does not scale linearly with respect to efficiency.