Optimizing Algorithm for Allocating Passengers in Shared Taxis

The issue of sharing vehicles has been riding since the '70s, but the advent of smartphones has made it a competitive choice to other transportation modes in recent years. The lack of restrictions on the movement of Internet-based passenger sharing systems leads to patrolling numerous personal vehicles in the network; this exacerbates congestion in high-traffic areas. On the other hand, the significant presence of circulating taxis and their non-optimal performance have disrupted the normal flow of traffic during peak hours and have led to an increase in travel time. This paper outlines a novel optimization algorithm for sharing repetitive and pre-planned trips. This algorithm is implemented on the midtown area network of Manhattan, New York, USA. Three scenarios were defined to simulate common services' status with the base scenario (do-nothing), which makes comparing possible with indicators such as distance travelled, and taxi occupancy ratio determined by passenger coefficient. Results of the first scenario - sending the nearest car - shows a decrease of 10.51%, the second scenario - allocating passengers to the nearest taxi - shows an increase of 10.16%, and finally the third scenario - the proposed algorithm - shows an increase of 25.56% in total mileage compared to the base scenario. Moreover, by defining Sharing Importance Factor (SIF) and using the proposed algorithm, it is possible to organize round-trip taxis, service repetitive and pre-planned trips, and significantly reduce the distance travelled throughout the network, and finally increase the passenger coefficient.