A hybrid approach to matching taxis and customers

We propose a Hybrid method for a computationally efficient matching between taxis and customers in a Mobility on Demand System. Our framework explores a hybrid path search for the assignment in order to minimize the total travel cost with a limited knowledge of the network. To validate the Hybrid method, two assignment techniques are compared: Greedy and Bipartite. The Hybrid method is a two stage process, where we first use the Euclidean distance-based travel cost for fast pairing between the taxis and customers. In the second stage, we iteratively unfold the street distance-based travel cost for the least cost path in the Euclidean metric. The outcome of the proposed framework is a fast and efficient taxi to customer assignment with the optimal total travel cost.