Application Of Intelligent Algorithm On Dynamic Flight Landing Scheduling

As one of the contents of Air Traffic Flow Management in terminal area, Aircraft Landing Scheduling is to reduce delay and controller's workload by the flight time and sequence optimization in zone control area and approach terminal control area. It has important theoretical value and practical significance that research flight landing scheduling algorithm and develop computer aid decision system of arrival flight scheduling in airport terminal area.The problem of flight landing scheduling is typical NP-hard problem. FCFS is a common method to solve the problem, but hard to give a reasonable scheduling scheme in the busier period. Intelligent algorithm has a strong ability of parallel and global search and has become a mainstream method of solving the problem. While in previous studies all of flights in airport terminal area as a whole are scheduled, it is difficult to be applied to real environment. Meanwhile, the studies almost not consider algorithm stability and real-time capability, and it is hard to be applied to the dynamic flight landing scheduling.According to the actual landing process of the aircrafts into terminal area, establish respectively the scheduling models of before and after entering initial approach point. The model of before entering initial approach point immediately adjust the speed to expand flights' level spacing by controlling acceleration. The other model is a segmentation model that flight approach process is parted and the case in the cross point is handled. According to the problem of Intelligent Algorithms difficult to obtain stable sorting result, two improved algorithms are used. One is based on the MPS constraint genetic algorithm. The other is a discrete particle swarm optimization algorithm. Finally, a small simulation system of flight landing in airport terminal area is built according to the two group models and improving the intelligent algorithms. The results show that improved genetic algorithm and immune particle swarm algorithm is better than FCFS method, and have a certain stability and real-time capability.