Research On Detection And Propagation Prediction Method For Arrival And Departure Flight Delays Based On Immune Algorithm

With the rapid development of domestic civil aviation industry, flight delaysbecome more and more serious, and especially for large busy airports, flight delayshave become a chronic illness. Therefore, the detection and prediction of flight delayshas become an urgent study and solve problems. The detection of the state of flightdelays and distinguishes between self and non-self in the immune system is greatlysimilar. So, from the view point of airports, based on immune algorithm andcombining the consideration of important factors for affecting the airdrome flightdelays, the paper would make real-time prediction for flights delay, around arrivalflights and departure flights. The main research work of this dissertation is as follows:In order to improve the accuracy of prediction of flights delay, this paperproposed an forecasting model for flight delays based on fuzzy immunization strategy,and combining immune detection and the thought of fuzzy membership grade, afunction of antibody detection for the set of flights delay was established. In theprediction process, real-time monitor changes of the environment, adaptively changethe confidence level, and select the detectors at the current confidence level.A largenumber of experimental results on domestic hub airports show that the model is moreadaptive and real-time. Departure flight delay is closely related to arrival flight delayand a large number of statistical analysis shows that in the same arrival delay state,departure delays will be approximately consistent with the Gaussian distribution.Therefore combining the immune operator and the Gaussian operator we proposed animmune detecting method for arrival flights and departure flights based on Gaussiandistribution. Since the arrival flight and departure flight formed a special flight chainand arrival delay, a delay caused by the arrival flight can transfer to the departureflight, we proposed a transfer analysis method for arrival and departure flights basedon Double-layers Markov model in which delay transfer between arrival anddeparture flights was considered a process of state transfer and the upper Markovchain fitted the state transition of the airport and the lower Markov chain fitted thestate transition of arrival delay and departure delay. The simulation results showedthat the algorithm is effective.