Generalized Flight Delay Prediction Method Using Gradient Boosting Decision Tree

Flight delay is a major issue in the aviation industry. In commercial aviation, if a flight reaches its destination 15 minutes later than the scheduled arrival, it is said to be delayed. Flight delays cause a great deal of bother to travelers. It could make them late to their booked occasions or miss a corresponding flight, accordingly prompting outrage and dissatisfaction. Likewise, travelers may not generally be entitled for a refund when a postponement happens. Carriers report that couple of the numerous reasons prompting most flight delays are carrier glitches, climate conditions, support issues with the airplane and congestion in air traffic. Rapid development in airline industry has led to an increased number of aircrafts in the skies, this has brought about air-gridlock causing flight delays. Flight delays are not only extremely undesirable financially but also have adverse environmental effects. Air-traffic management is becoming increasingly challenging. The aim of our research work is to predict the delay of flights due to various factors using machine learning and deep learning so as to minimize losses and increase customer satisfaction. This Machine Learning model could be integrated with Airlines systems for the use of staff and customers and it could also rank Airlines and flights based on delays.

Flight delays have become a major issue in the aviation industry, impacting operational efficiency and customer satisfaction. This study proposes a CatBoostClassifier-based approach combined with Feature Engineering, Bayesian Optimization, and Random Over Sampling techniques to improve the accuracy of flight delay predictions. Based on model evaluation results, the use of Feature Engineering and Bayesian Optimization enhances performance compared to the baseline CatBoost model. The CatBoost+FE+Bayes combination achieves an accuracy of 83.32%, higher than the unmodified CatBoost model, which only reaches 82.95%. However, applying the Random Over Sampling technique in the CatBoost+FE+Bayes+ROS combination decreases model performance, reducing accuracy to 81.44%. Regarding other metrics, the CatBoost+FE+Bayes model demonstrates the highest F1-score of 0.62, indicating a balance between precision and recall. Additionally, the Area Under Curve (AUC) analysis reveals that CatBoost+FE+Bayes has the highest AUC value of 0.7793, followed by CatBoost+FE at 0.7768, and the unmodified CatBoost model at 0.7643. Meanwhile, the application of ROS leads to a decrease in AUC value to 0.6787. These findings suggest that utilizing Feature Engineering and Bayesian Optimization significantly enhances flight delay predictions. However, resampling techniques such as ROS do not always positively impact the tested model and can even degrade classification performance. The objective of this research is to develop a more accurate flight delay prediction model through the application of appropriate optimization techniques. The resulting model is expected to improve prediction quality and benefit the aviation industry by optimizing operational efficiency and minimizing the negative impact of delays on passengers.

The purpose of this paper is to combine the cloud model method with the Artificial neural network to predict and evaluate flight delay date. In this paper, the flight delay data of the United States in recent 15 years were collected, involving nine airlines and seven delay causes. Then, the flight delay prediction model is obtained by combining neural network and cloud model theory. 36 sets of data are extracted to verify the accuracy of the prediction model. Finally, the delay rate is predicted and the prediction data is evaluated by cloud model. The results show that the amount of data to be processed is reduced to 20 % after combining the cloud model with the neural network. In the model validation, the average expected value deviation of flight delay prediction model is 23.1 %. In the analysis and evaluation of the prediction results, the delay cause 6 is taken as an example, with the highest delay rate of 7.03 % in December. The delay risk assessment results for all months were general. The method provided in this paper can predict flight delays and provide theoretical support for the delay prevention of airlines.

To accommodate the unprecedented increase of commercial airlines over the next ten years, the Next Generation Air Transportation System (NextGen) has been implemented in the USA that records large-scale Air Traffic Management (ATM) data to make air travel safer, more efficient, and more economical. A key role of collaborative decision making for air traffic scheduling and airspace resource management is the accurate prediction of flight delay. There has been a lot of attempts to apply data-driven methods such as machine learning to forecast flight delay situation using air traffic data of departures and arrivals. However, most of them omit en-route spatial information of airlines and temporal correlation between serial flights which results in inaccuracy prediction. In this paper, we present a novel aviation delay prediction system based on stacked Long Short-Term Memory (LSTM) networks for commercial flights. The system learns from historical trajectories from automatic dependent surveillance-broadcast (ADS-B) messages and uses the correlative geolocations to collect indispensable features such as climatic elements, air traffic, airspace, and human factors data along posterior routes. These features are integrated and then are fed into our proposed regression model. The latent spatio-temporal patterns of data are abstracted and learned in the LSTM architecture. Compared with previous schemes, our approach is demonstrated to be more robust and accurate for large hub airports.

Nowadays, the civil aviation industry has a high precision demand of flight delay prediction. To make full use of the characteristics of flight data and meteorological data, two flight delay prediction models using deep convolution neural network based on fusion of meteorological data are proposed in this paper. One is DCNN (Dual-channel Convolutional Neural Network), which refers to the ResNet network structure. The other is SE-DenseNet (Squeeze and Excitation-Densely Connected Convolutional Network), combining the advantages of DenseNet and SENet. Firstly, flight data and meteorological data are fused in the model. Then, both DCNN and SE-DenseNet models are used to extract feature automatically based on the fused flight data set. Finally, the softmax classifier is adopted to predict the flight delay level. For proposed DCNN model, both straight channel and convolution channel are designed to guarantee the lossless transmission of the feature matrix and enhance the patency of the deep network. For proposed SE-DenseNet model, a SE module is added after the convolution layer of each DenseNet block, which can not only enhance the transmission of deep information but also achieve feature recalibration in the feature extraction process. The research results indicate that after considering characteristics of meteorological information, the accuracy of the model can be improved 1% compared with only considering the flight information. The two deep convolutional neural networks proposed in this paper, DCNN and SE-DenseNet, can both effectively improve the prediction accuracies, reaching to 92.1% and 93.19%, respectively.

By predicting flight delays, economic losses caused by delay situations can be effectively prevented and reduced. A classification algorithm for flight delay prediction based on unbalanced data is proposed to solve the problem of unbalanced number of delayed and non-delayed samples in the process of flight delay prediction. Firstly, a multi-classification model based on the CycleMLP algorithm is proposed for flight delay prediction. Secondly, the CycleMLP algorithm is improved by using the Focal Loss modified cross-entropy loss function to build the final classification model for flight delay prediction. And the experimental results show that the improved model has better prediction results for delayed flights, with a 5.54% increase in average macro precision and 1.86% increase in accuracy. Finally, the flight delay prediction results are published in the mobile APP application system to provide corresponding advice to passengers and airport departments.

Flight delay is a significant challenge faced by air traffic control systems, severely affecting passenger travel experience and airline operational efficiency. This paper addresses this issue by proposing an intelligent flight sorting method based on machine learning and deep learning. By preprocessing and feature engineering of flight historical data, airport data, and route data, a set of high-quality feature sets is constructed. On this basis, models such as random forests and LSTM are employed to predict flight delays and intelligently sort flights according to the prediction results. Experimental results demonstrate that this method significantly enhances the accuracy and efficiency of flight sorting, reducing the impact of flight delays on passenger travel. This study provides a novel and effective approach to addressing flight delay problems, offering broad application prospects and practical value.

PurposeThis paper aims to reduce flight delay propagation, improve flight punctuality rate and ensure aircraft maintenance opportunities by establishing an integrated aircraft scheduling model, aiming at minimizing the total propagated delay and direction operational cost.Design/methodology/approachIn this paper, flight data sets are obtained through automatic dependent detection broadcast. To accurately predict flight delay time, the flight delay prediction eXtreme gradient boosting model adds the data set obtained by random forest advance model learning and predicts the newly generated flight delays. Finally, based on the forecast results, the flight plan can be optimized and adjusted by using the improved column generation algorithm.FindingsIt is verified by the actual weekly planned operation data of an airline company, experiments show that the model established in this paper can reduce flight delay propagation by 30% in case tests and each aircraft has the opportunity to be repaired at the base airport.Originality/valueOptimize the aircraft scheduling plan, cover a wide range of data, not just a single route and airport, supplement the gap in the aircraft scheduling plan based on weather factors to predict flight delays.

Delays in flights and other airline operations have significant consequences in quality of service, operational costs, and customer satisfaction. Therefore, it is important to predict the occurrence of delays and take necessary actions accordingly. In this study, we addressed the flight delay prediction problem from a supervised machine learning perspective. Using a real-world airline operations dataset provided by a leading airline company, we identified optimum dataset features for optimum prediction accuracy. In addition, we trained and tested 11 machine learning models on the datasets that we created from the original dataset via feature selection and transformation. CART and KNN showed consistently good performance in almost all cases achieving 0.816 and 0.807 F-Scores respectively. Similarly, GBM, XGB, and LGBM showed very good performance in most of the cases, achieving F-Scores around 0.810.

Aiming at the phenomenon of a large number of flight delays in the terminal area makes a reasonable scheduling for the approach and departure flights, which will minimize flight delay losses and improve runway utilization. This paper considered factors such as operating conditions and safety interval of multi runways; the maximum throughput and minimum flight delay losses as well as robustness were taken as objective functions; the model of optimization scheduling of approach and departure flights was established. Finally, the genetic algorithm was introduced to solve the model. The results showed that, in the program whose advance is not counted as a loss, its runway throughput is improved by 18.4%, the delay losses are reduced by 85.8%, and the robustness is increased by 20% compared with the results of FCFS (first come first served) algorithm, while, compared with the program whose advance is counted as a loss, the runway throughput is improved by 15.16%, flight delay losses are decreased by 75.64%, and the robustness is also increased by 20%. The algorithm can improve the efficiency and reduce delay losses effectively and reduce the workload of controllers, thereby improving economic results.

In day-to-day operations, flight delays might happen due to some reason such as bad weather conditions, equipment failure, and limited runway capacity. Flight delays do not only increase crew operating cost, they can cause delay propagation on subsequent flights. Since the crew cost is the second largest cost, it is important to construct a robust flight schedule where any flight delay will not lead to a massive disruption to the subsequent downstream flights. In this paper, we analyze the effect of flight retiming to the robustness of flight schedule. We construct a simulation for measuring the robustness of the flight schedule, where the probability of crew pairing found after a number of flight delays happen is used as the measure of robustness. It is shown that the retiming will increase the robustness of the schedule if we only retime the flights which appear in the first cycle of the set of selected pairings.

Flight delays are unavoidable, and they have a substantial financial impact on airlines. Airlines demand an accurate flight delay estimate since the data may be used to improve customer satisfaction and profitability. Many research have been conducted on flight delay modeling and prediction, with the majority of them attempting to predict the delay by extracting important qualities and variables. Due to the large volume of data, dependencies, and large number of parameters, most of the offered solutions are insufficiently accurate. This research presents a Deep Learning-based approach for anticipating aircraft delays (DL).DL is a relatively recent method for tackling issues with a high level of complexity and a large number of data. Besides, DL can extricate key highlights from information naturally. In expansion, since most flight delay information is boisterous, a technique based on stack denoising autoencoder is created and included within the proposed model. Furthermore, the Levenberg-Marquart calculation is utilized to decide weight and inclination fitting values, and the yield is at that point tuned to convey exceptionally precise comes about. Two more structures are created to investigate the effect of the stack denoising autoencoder and the LM method on the model structure.

Data-Driven Robust Aircraft Assignment to Minimize Delay Propagation

Integrated development and operational collaboration of regional airport groups have the potential to improve capacity and safety and also reduce environmental impacts and operational costs. However, research in multiairport systems (MASs), especially in China, is still in its infancy, with the consequences of unbalanced development, inadequate coordination, unclear function partitioning, difficulty in air traffic management, and poor service quality of regional airports. Considering these characteristics influencing effective interaction and collaboration of regional airports, this paper formulates a model to optimize the flight schedules in the MAS with multiple objectives of minimizing the maximum displacement of all flights, the weighted sum of total flight adjustment of each airport, and flight delays. An improved simulated annealing algorithm (SAA) is designed to solve the proposed multiobjective optimization problem. The model is applied to a case of the Beijing-Tianjin-Hebei Airport Group. The computational results demonstrate that the model generates significant reductions in maximum displacement, average displacement, and average delay, compared to the First-Come-First-Served (FCFS) principle. The model proposed in this paper can be used by civil aviation authorities, air navigation service providers, and airlines to facilitate the integrated management of flight schedules in the MAS.

The extended cellular automata (ECA) is one of the extending application of cellular automata (CA) by extending the component of cell and the definition of cellular neighbors. Based on ECA, the definition of parallel cellular automata (PCA) composed of multi-ECA is proposed and the evolutions of all ECAs are parallel. Depending on the model of PCA, it becomes more easily to describe the synchronous evolution and the state propagation of a complicated system, which is composed of multi-subsystem. Flight delay and delay propagation among multi-aerodrome is one of the problems encountered by most airlines. The simulation based on the model of PCA shows that the model of PCA is a new approach to describe the prediction of flight delay and delay propagation.