Ridesharing user equilibrium model without the en-route transfer: An OD-based link-node formulation

As an emerging and sustainable mode of traffic, the widespread ridesharing programs can fully utilize existing vehicle resources and reshape traffic flow distribution. The study aims to explore the spatial distribution of mode-split traffic flow and appraise the impact of ridesharing services on urban traffic conditions. To achieve these goals, by constructing a three-layer extended network to visually depict mode choices and employing the link-node modeling approach, a ridesharing user equilibrium (RUE) model in terms of OD-based link flows is first established to formulate travelers’ mode and route choice behavior in the traffic network with ridesharing programs, which avoids the path enumeration and storage required by path-link modeling approaches. By relaxing the same origin and same destination assumption, the RUE model is reformulated to suit a more general scenario, where ridesharing drivers can pick up multiple passengers along their routes. Compared with other link-node modeling approaches in terms of origin-based or destination-based link flows, the unreasonable phenomenon that a ridesharing passenger needs to sequentially ride multiple ridesharing vehicles for a single trip (i.e., en-route transfers) is avoided. Moreover, we prove the conditions that the path flow of travelers can be derived from the OD-based link flow, which implies that we can know the travelers’ path selection under some mild conditions. Finally, several numerical experiments are conducted on three transportation networks at different scales to analyze the model performance.