Flight Pushback Control And Optimization Research Based On Departure Taxiing Time Prediction

With the rapid economic and social development,the demand of air transportation is increasing,but the low airport field operation efficiency and frequent flight delays expose the contradiction between airport resource allocation and aviation demand.Due to the large investment and long period of airport renovation and expansion,it is impossible to alleviate the current situation of airport traffic in the short term.Therefore,it is of great practical significance to carry out advanced and scientific research on airport operation management strategies to alleviate airport congestion and improve airport operation efficiency.Firstly,the main factors affecting the departure taxiing time of flights are analyzed by combining the airport infrastructure and inbound and outbound operation processes.The data of selected influencing factors are extracted from the actual operation data as the prediction model data features,and the feature data are analyzed from the perspectives of data distribution,missing value processing,text feature processing and correlation.Considering the problem of correlation between features and features,the data are decoupled by using principal component analysis to obtain the final data of the prediction model.Secondly,considering the difficulty of adjusting the parameters of the XGBoost(Extreme Gradient Boosting)model,the GWO(Grey Wolf Optimizer)algorithm is introduced to find the optimal combination of parameters.To overcome the shortcomings of the GWO algorithm,the convergence factor and the position update formula are improved,and finally the IGWOXGBoost taxiing time prediction model is built.Finally,the relationship between departure taxiing flight volume,take-off flight volume and pushback flight volume is analyzed,the state transfer equation of departure taxiing flight is established,and the relationship between departure taxiing flight volume,take-off flight volume and arrival flight volume is analyzed,the departure rate saturation curve is plotted based on historical data,the critical departure taxiing flight volume under different arrival volumes is derived,the pushback rate control method is proposed in combination with the taxiing time prediction.The pushback time slot allocation optimization model is established with the minimum flights departure cost,and the firefly algorithm is designed to solve the model,and the feasibility and advancement of the pushback control method and the time slot allocation optimization model are verified by example analysis.The research results show that the simultaneous inbound flights volume,simultaneous departing flights volume and queue length show similar distribution,the simultaneous inbound flights volume,simultaneous departing flights volume and queue length have a relatively strong correlation,and the queue length has a relatively strong correlation with the average departing taxiing time of the preceding flights,and the average departing taxiing time of the preceding flights,and the queue length also has a high correlation with the departing taxiing time.Compared with the three prediction models of GA-XGBoost,BP Neural Network and SVR(Support Vector Regression),the IGWO-XGBoost model established in this paper has higher prediction accuracy and smaller error fluctuations,with an accuracy of 96.7%within 5min error.After the implementation of the pushback rate control,the average departure taxiing flight per time window was decreased from 8.25 to 7.13,a decrease of13.6%,and the average take-off rate was improved from 5.38 to 5.88,an improvement of9.3%.The implementation of the pushback rate control method has improved the operational efficiency of the airport field and reduced airport congestion.Compared with the original pushback strategy,the average flight parking space waiting time of the optimized pushback strategy increased from 4.8min to 6min,an increase of 26.0%;the departure taxiing time decreased from 14.6min to 13.3min,a decrease of 8.9%,and the total flight departure cost decreased from 112,148.3RMB to 98918.4RMB,a decrease of 11.7%.The optimized pushback strategy converts the flight waiting time at the runway end into the parking space waiting time,which reduces the departure cost and improves the airport field operation efficiency.The research results of this paper can provide references for airport management to develop efficient flight pushback plans and scientific and reasonable field scheduling.