A Two-Phase Route Optimization Model For The Hybrid Demand Responsive Airport Shuttle

With the improvement of travel quality and the continuous increase of air passenger traffic,the service capacity of the airport’s landside arrival and departure system is facing new challenges and opportunities.At present,the airport landside transportation dominated by car deviates from the requirements of development under the dual-carbon policy,and cannot provide support for the construction of a transportation powerhouse with green transportation as one of the core orientations.Therefore,how to improve the attractiveness of airport landside public transportation represented by airport shuttle has become an important issue in the development of air transportation in the next stage.Facing the differentiated travel demand,this thesis proposes a new service mode of hybrid demand responsive airport shuttle by integrating the reservation and real-time modes,and constructs a two-phase route optimization model of the hybrid demand responsive airport shuttle.The arrival and departure system of the Capital International Airport is used as a case to analyze the application.The research results will provide an important reference for the formulation of strategies to improve the operating efficiency and service quality of airport shuttle,as well as provide supports for the method to improve the share airport landside public transportation.The specific research is as follows.Firstly,the operation characteristics and service advantages of the hybrid demand responsive airport shuttle are clarified.The spatiotemporal dynamic pricing strategy,station selection and service strategy are designed.The location selection and service strategy of fixed stations are based on the K-MEANS algorithm combined with the improved Hierarchical Clustering.In this algorithm,the spatiotemporal similarity measure of reservation passenger is used to normalize the spatiotemporal information.The location strategy of fixed stations and its service time are output based on the algorithm.Demand responsive stations are based on service units and a batch dynamic insertion strategy is developed,which simplifies computing and reduces the impact on fixed station services.Secondly,a two-phase route optimization model and design algorithm are developed.In the first phase,the initial route optimization model is constructed according to reservation demand.The multi-objective of the initial route optimization model are to minimize the cost of the enterprise and the time cost of passenger.The direct cost of enterprise in the objective function includes the fixed cost of the shuttle,the variable cost of the shuttle.The potential cost of enterprise includes the cost to reject passenger request.The revenue of the enterprise includes dynamic fares for reservation passenger.The direct cost of passenger in the objective function includes the walking time and the time in the car.The potential cost of passenger includes passenger service deviation time.The NSGA-II algorithm is designed to solve the multi-objective optimization model and output the Pareto solution set including the station service time and the running route.In the second phase,the dynamic route optimization model is constructed according to the real-time demand of passenger.The objective of the dynamic route optimization model is to minimize the total cost of enterprise and passenger.The direct costs include the additional operating costs of vehicles,the penalty cost of slack time.The potential costs include the cost to reject passenger requests and detour time costs for passenger.The revenue of the enterprise includes dynamic fares for real-time demand of passenger.The genetic algorithm is designed to solve the dynamic running route of the vehicle.Finally,the temporal and spatial characteristics of landside passenger flow at the Capital International Airport are analyzed.Taking Beijing Huitian area to the Capital International Airport as an example,fixed stations are designed based on the station location service algorithm,and the improved algorithm is compared with the output results of the traditional clustering algorithm.The reliability of the improved algorithm is verified.In the first phase,based on the Pareto solution set output by the first-phase route optimization model,the lowest enterprise cost plan,the lowest passenger time cost plan,and the comprehensive plan are analysed.The reliability of the initial route optimization model is verified by sensitivity analysis using parameters such as the number of vehicles and the unit fixed cost of vehicle.In the second phase,the second-phase route optimization model is used to output the dynamic route.The effectiveness and robustness of the optimization model are proved by the sensitivity analysis of the penalty cost per unit of slack time and the weight value of different goals.