Research On Flight Delay Propagation Expanded Prediction By Step Based On Uncertain Information

In recent years, with the increasing number of flight transport, the flight delay problem is becoming more and more serious. Flight delays not only cause serious economic losses to the airlines, but also disturb the normal production order of civil aviation. Now that an aircraft need to perform several flights and reach to several airports one day, the delay of upstream flights to downstream flights have many uncertainties, so how to predict downstream flights delay level based on the upstream flight delay status is a challenge. The paper research on flight delay propagation expanded prediction by step based on uncertain information is based on this background. Following are the main works:Firstly, the operation mechanism of flight and the formation of flight delays are illustrated. Secondly, we make models about turnaround time, which is the key of flight delay transform. The regression trees used to predict the turnaround time and the performances of models on different parameters are compared. Bayesian network is also applied to estimate the turnaround time. Dynamic estimation method about the turnaround time is proposed based on the Bayesian network parameter incremental learning method. Sensitivity analysis toward the factors is made to compare effects of them. Then we compare the results of two models and explain the explain advantages of Bayesian network. Thirdly, we make study about flight delay models from the perspective of flight chain. To single airport, we built the arrival and departure model.To transformof the flight chain, we put forward the expanded prediction of flight delay model. The characteristics of flight delay transmission are analyzed based on Dynamic Bayesian network. The result indicates that uncertainty about delay time is increasing when the delay is transmitting so the possibility of accurate prediction is decreasing. Finally, the prediction system of flight delay transmission is introduced.The research outcomes have high value to the warning and decision support to the frequent flight delay of China.