A Fast Response Multi-Objective Matching Algorithm for Ridesharing

In many metropolitans, especially during rush hours on holidays, thousands of riders will initiate travel orders at the same time, and the existing carpool matching model cannot handle largescale travel orders quickly enough. For handling this problem, a fast and efficient multi-objective carpool matching algorithm (MOCMA) is put forward, which generates a set of different matching schemes suitable for different practical scenarios. First, the idea of partition is adopted to gather riders and drivers with similar journeys, and the relationship matrix construction algorithm (RMCA) is proposed; then from the perspective of riders and drivers, the maximum service quality and the maximum shared mileage are two objectives, and a set of non-dominated solution sets are generated using MOCMA; finally, the simulation experiment results show that MOCMA proposed is suitable for different practical scenarios, the matching success rate is as high as 99.7%, and it has significant advantages over MOEA/D, SPEA2, and FastPGA.