Taxi-Pooling Scheduling Model and Algorithm Based on Many-to-Many Pickup and Delivery Problems

With the development of mobile network technology, many kinds of taxi systems based on application (APP) are appearing, which provides convenience for travelers and makes taxi-pooling possible. But the taxi systems are working almost based on independent choices of parties, drivers and passengers, and some systems are not efficient enough. This easily leads to defaults. Particularly, order-competing always influences driving safety. In this paper, we divide one day into several periods, the dynamic taxi scheduling problem is converted into a static one. A taxi-pooling scheduling model is built based on many-to-many pickup-and-delivery problems (MPDPs), in which the information such as taxi positions and passengers’ demands are known, travel time, and passengers cost are considered. An intelligent heuristic algorithm is designed, and we get the optimization results of two instances which included passenger distribution program, taxi-pooling plan, and taxi routes scheme. Our method is efficient, especially in a large-scale road network.