The Simulation of Dynamic Shared Taxi Routing Problem

Abstract

In recent years, taxi sharing is becoming a popular transportation mode. Compared with the traditional operation mode, taxi sharing possesses some advantages. For example, it allows several passengers to share one taxi in the same route thus improving the utilization rate of taxi seats and reduces the taxi traveled miles. Besides, taxi sharing decreases the number of taxis needed, which efficiently relieve the traffic congestion and reduces the emission of carbon dioxide. In this paper, we consider the dynamic shared taxi routing problem (D-STRP) and develop a simulator to mimic the process of taxi sharing. To maximize the objective of the proposed problem, we define a parameterized taxi allocation policy to make a trade-off between the passenger wait and detour time and taxi route length. Furthermore, we propose the insertion algorithm to assign the passenger request to the best available at the lowest expense. The simulation results indicate that the parameterized taxi allocation policy has obvious impact on the objective of D-STRP. Through simulations, we can find the optimal taxi allocation policy in each trip scenario fast, which is of great significance for the taxi service provider.