Research On Airport Flight Flow Prediction Based On Cluster Analysis And Neural Network

With the rapid development of civil aviation in China,airport flight traffic is increasing day by day.The limited runway capacity and airspace resources are increasingly unable to meet the increasing traffic demand,which leads to a large number of flight delays.The purpose of this paper is to predict future changes in airport flight flows based on machine learning and data mining techniques and using historical data.This study helps to better grasp the changes of airport flight flow,ensure the flight order over the airport,and relieve the congestion of aircraft routes over the airport in order to improve the operational efficiency of the airport terminal area,as well as to guarantee the safety of flights and improve the efficiency of air transportation companies.By analyzing historical data,we can predict the number of future flights,which helps airport managers to make reasonable flight scheduling plans to ensure the normal operation of flights and reduce delays and congestion.In this thesis,we combine neural network and cluster analysis to predict the traffic in the terminal area of the airport,so as to improve the accuracy of traffic prediction.Firstly,we analyze the characteristics of air traffic flow in the terminal area,then focus on the basic preparation work such as collection,quantification,cleaning and fusion of flight operation and aviation meteorological data,correlation analysis of data,elimination of redundant elements,and extraction of key elements that affect the flow.Then,the daily feature vectors required for clustering are extracted based on the obtained key elements,and the similar day clustering analysis is performed based on the K-means algorithm to build a traffic similar day database.To solve the difficult problem of hyperparameter selection for neural networks,a chaotic particle swarm optimization algorithm based on adaptive inertia weights is proposed.The algorithm is improved and optimized on the basis of the basic particle swarm algorithm,and a new dynamic adaptive inertia weight is used to strike a balance between global and local optimization,and then the chaos idea is combined with the particle swarm algorithm to solve the problem that the particle swarm algorithm tends to fall into local optimum.Finally,the clustering analysis and AC-Bi LSTM method are combined to predict the traffic in the terminal area.Based on the construction of the AC-Bi LSTM short-term traffic prediction model,the ACBi LSTM model combining cluster analysis and hyperparameter optimization will be used for short-term prediction of the terminal area traffic.Finally,the effectiveness and accuracy of the method are verified by examples.