With the development of aviation industry and the improvement of people's living standards, travelling by air has become more and more popular in China, which puts forward higher requirements to airport landside transportation. Amongst others, the airport bus, due to its advantages such as affordable, large capacity, energy-saving and so on, has become an important mode for carrying people from and to airports. In this paper, we investigate an airport bus scheduling problem to better meet people’s daily travel demand. More specifically, by analyzing the data from car-hailing and taxis, potential passenger volume of starting a new airport bus line is predicted. Next, by applying the K-Means clustering analysis, appropriate sites are determined based on the distribution of the passenger volume, and the most suitable route is selected by comparing the time cost and bus cost of passengers among different routes with the local search method. Afterwards, a multi-objective operation and scheduling model is developed which takes the benefits of both passengers and enterprises into account, and the departure interval of the airport bus is estimated by adopting the improved genetic algorithm. The model is validated by a case study using data from the Beijing capital international airport.

CICTP;Genetic Algorithm (GA);Bus scheduling;Airport