An airport shuttle bus (ASB), as an environmentally friendly mode of green transportation, is an effective way to solve the "first/last mile" of aviation passengers, which can attract a higher passenger transfer from private cars to public transport, thereby reducing emissions of carbon dioxide and other polluting gases. This study presents a multi-objective mixed-integer linear programming for ASB services in a dynamic environment. Taking into account time-varying demand and travel time characteristics in different periods, the proposed model provides a comprehensive framework that simultaneously advises passengers to join the bus at the nearest bus stations, designs routes for transporting them from these selected stations through the airport, and computes their departure frequencies in multiple periods. The primary objective is to optimize both the total ride time and waiting time for all passengers. The secondary objective is to optimize the total transfer distance of all passengers simultaneously. Given the Non-Deterministic Polynomial (NP) hardness of this problem, a two-stage multi-objective heuristic approach based on the non-dominated sorting genetic algorithm (NSGA-II) is combined with a dynamic programming search method and further advanced to obtain the Pareto-optimal solutions of the proposed model within a reasonable time. Finally, the proposed model and algorithm feasibility are proved by a test example of designing a shuttle bus route and schedule at Tianjin Airport, China. The results show that the total passenger travel time of the presented model is markedly reduced by 1.21% compared with the conventional model.
Read full abstract