A Novel Framework for Ridesharing Services

To lighten traffic overcrowding caused by the quick growth of vehicles in big cities, this study proposes a new framework for ridesharing services. Our goal is to serve as many requests for travel as possible in the meantime minimizing the total travel distance of the driver. Towards this goal, we employ techniques for analyzing historical data in service decision making. The process of finding groups for ridesharing based on the practical observation that many people travel for a long distance, head in the same direction during the same certain interval of time every day. The mining algorithm is applied to discover frequent routes of a user. The frequent routes of users then can be used to construct rideshare groups in advance which help reduce waiting time for the regular users. The ridesharing model enables users to join a rideshare group in either offline or online fashion. The schedule of each group is generated with the shortest distance. A user can join a group in real-time if spatial and temporal constraints are satisfied.