The flex-route transit service routing plan considering heterogeneous requests and time windows

Abstract

Flex-route transit (FRT) service has become increasingly popular in low-density suburban areas as a shared transportation mode. Improving the service quality of FRT heavily relies on effective routing plans. To end of this, this paper addresses the FRT routing plan problem considering heterogeneous requests and time windows (FRTRPP-HRTW), which is formulated as a mixed-integer programming (MIP) model. The MIP model concerns on the objective of minimizing the average system costs. Due to the NP-hard nature of the considering problem, a niche-based modified ant colony optimization (NM-ACO) algorithm is developed to handle it efficiently for large-scale cases. In NM-ACO, a niche-based pheromone update technology is designed to increase the population diversity. Finally, as a case study, a real-world FRT service is employed to validate the effectiveness of NM-ACO. Computational results demonstrate that the NM-ACO can obtain high-quality solutions with small gap in much shorter time compared with the mature solver (i.e. CPLEX). Furthermore, compared with the well-known MMAS algorithm and a state-of-art ACO-DI algorithm, the NM-ACO expresses higher precision and stronger robustness.