Analytical and Agent-Based Model to Evaluate Ride-Pooling Impact Factors

Abstract

On-demand ride-pooling (ODRP) services have the potential to improve traffic conditions in cities and at the same time offer user-centric mobility services. Recently, an analytical model, which investigates the influence of service quality parameters, such as detour, maximum waiting time, and boarding time, on the fraction of trips which could potentially be shared (a quantity called shareability), has been presented. The aim of this study is to test this model with a simulation framework that models an ODRP service in different levels of detail. The results show that by increasing the modeling complexity, in which we consider network topology, trip distribution patterns, optimization objectives, and changing velocity, the theoretical value of shareability and the actual experienced shared rides are decreased. It is observed that the shareability predicted by the mathematical model could be confirmed by a certain simulation setup with the objective to maximize shared rides. Nevertheless, changing the optimization objective to optimizing the total kilometers driven has the highest impact on shareability, decreasing it by up to 50%. By using a fitting procedure within this simulation setup, we can still exploit the analytical model to predict the influence of service quality parameters. This study may be useful for other researchers who plan to model ride-pooling systems and for operators who want to have an estimation of the level of shared rides they can achieve in an operating area.