Ride Sharing Using Dynamic Rebalancing with PSO Clustring A Case Study of NYC

Abstract

The shared vehicle can improve the efficiencyof urban mobility by reducing car ownership and parkingdemand. Existing rebalancing research divides thesystem coverage area into defined geographical zones,but this is achieved statically at system design time,limiting the system’s adaptability to evolve. In the currentstudy, a method has been proposed for rebalancingunoccupied vehicles in real-time while considering travelrequests, using a bio-inspired method known as ParticleSwarm Optimization clustering (PSO-Clustering). Thesolution was examined using data on taxi usage inNew York City, first looking at the traditional system (noridesharing, no rebalancing), then carpooling, and finallyof both ridesharing and rebalancing.