The culture of sharing instead of ownership is sharply increasing in individuals behaviors. Particularly in transportation, concepts of sharing a ride in either carpooling or ridesharing have been recently adopted. An efficient optimization approach to match passengers in real-time is the core of any ridesharing system. In this paper, we model ridesharing as an online matching problem on general graphs such that passengers do not drive private cars and use shared taxis. We propose an optimization algorithm to solve it. The outlined algorithm calculates the optimal waiting time when a passenger arrives. This leads to a matching with minimal overall overheads while maximizing the number of partnerships. To evaluate the behavior of our algorithm, we used NYC taxi real-life data set. Results represent a substantial reduction in overall overheads.

The Merits of Sharing a Ride