关键词: 进离场航班排序, 跑道入侵, 不确定, 启发式算法

Abstract: According to the International Civil Aviation Organization (ICAO), a runway invasion is any incident at an airport that involves the mistaken appearance of aircraft, vehicles, and pedestrians on the surface of a protected area used for aircraft take-off and landing. Obviously, all kinds of intrusion events have the characteristics of randomness and suddenness, and there are differences in the place, time and duration of their occurrence. When a runway incursion occurs, aircraft are prohibited from taking off or landing during this time, which affects flight arrivals and departures scheduling problems (FADSP) for some flights. As can be seen from the above, FADSP with flight arrivals and departures scheduling problem invasion(FADSPI) is more complicated than the traditional FADSP, mainly in two aspects: (1)Considering that the priority of incoming flights is higher than that of departing flights, some runway incursions result in excessively long waiting times for aircraft on some flights, which need to land at alternate airports due to limited reserve fuel. (2)It is urgent to analyze the internal relationship among the randomness of intrusion events, the approach and departure sorting scheme and flight delay. Therefore, FADSPI helps to improve the scientific level of flight arrival and departure management in emergencies, thereby improving runway capacity and avoiding large-scale flight delays.
This paper proposes an optimization model for flight sequencing with multi-runway operation mode under uncertain scenarios. It is assumed that the probability of occurrence of each runway invasion event at different durations can be estimated, and the priority difference of delay and alternate handling for different flight types such as special aircraft, VIP and ordinary flights can be obtained in advance. The model tries to select some ordinary approach flights to alternate airports, allocate the remaining flights to different runways, and determine their take-off or landing time on corresponding runways, so as to reduce the delay cost of ordinary flights as far as possible under the circumstance of giving priority to special planes and incoming and outgoing passenger flights. According to the characteristics of the problem and based on the priority of inbound and outbound flights,a multi-stage parallel distributed heuristic algorithm is designed to solve the problem. Finally, a real case is used to analyze the difference in flight arrivals and departures ordering results under different invasion events. The relationship among the spatial and temporal distribution of intrusion events, the number of flights of different priority types and the delays is revealed to verify the correctness of the model.
The main findings are shown as follows: (1)When considering the priority of flights, although part of the runway slot resources are wasted and the arrival and departure delay time is slightly increased, the losses caused by VIP flight delays are reduced and the actual special needs for flight arrival and departure sequencing are met. (2)When supply (the number of runways) is less than demand (the number of inbound and outbound flights), the queue length of inbound and outbound flights will increase. Otherwise, the queue length will be reduced. Because flights arrive unevenly, the length of the queue fluctuates from moment to moment. Since overall demand is greater than supply, it takes more time for all the queues to dissipate. (3)Due to the different runway operation mode and the proportion of inbound and outbound flights, when the location, time and duration of runway invasion are different, the queuing formation and dissipation process of inbound and outbound flights will be significantly different. (4)When the runway invasion time increases, it may cause diversion flights to appear, and the number of diversion flights will increase with an increase in invasion time.

Key words: flight arrivals and departures order, runway incursion, uncertain, heuristic algorithm