Ridesharing user equilibrium model with drivers matching passengers from multiple OD pairs and avoiding the en-route transfers

This paper proposes a new path-based ridesharing user equilibrium model, in which a ridesharing driver share vehicles sequentially with passengers from multiple different origin-destination (OD) pairs during the entire journey. Simultaneously, each ridesharing passenger is picked up by only a ridesharing driver. Compared with existing ridesharing user equilibrium models, the proposed formulation eliminates the unreasonable phenomenon that en-route transfers. Two algorithms are developed to enhance computational efficiency: the first determines the maximum number of passenger OD pairs that a driver can accommodate, while the second identifies all feasible subpath combinations on the driver’s route. These algorithms enable the proposed model to handle large-scale networks effectively. Numerical experiments on the Braess, Nguyen-Dupuis, Sioux-Falls, and Chicago networks demonstrate the model’s capability to produce realistic user equilibrium flows with high computational efficiency.