Designing Optimization Algorithm of Taxi Circulation Using Predictive Approach: Chicago Case Study

Abstract

Commuting by taxi is one of the common modes of transportation in metropolitan areas. Taxi companies, however, have various distribution patterns and typically work separately from each other. Lack of coordination between these systems can potentially cause a waste of time and money. The primary purpose of this research is designing a new algorithm with an optimization core to minimize the total number of taxis in the system. This algorithm is remodeled for four different taxi allocation strategies, which are Earliest Ride, Closest Ride, Trip Feasibility Ratio (TFR), and Busiest Area. This research has assessed these models for Chicago trip data to investigate the impact of each model on the time and mileage efficiency of the system. The result of this research highlights that the Earliest Ride and Busiest Area models improve the total number of taxis. Also, the analysis of the number of one-ride taxis shows that the Busiest Area model reduces the number of one-ride taxis, specifically in Chicago’s high-demand community areas.