Deep Learning for Flight Demand Forecasting

Liya Wang,Amy Mykityshyn,Craig Johnson,Benjamin D. Marple

DOI: https://doi.org/10.48550/arXiv.2011.04476

2021-11-05

Abstract:Inspired by the success of deep learning (DL) in natural language processing (NLP), we applied cutting-edge DL techniques to predict flight departure demand in a strategic time horizon (4 hours or longer). This work was conducted in support of a MITRE-developed mobile application, Pacer, which displays predicted departure demand to general aviation (GA) flight operators so they can have better situation awareness of the potential for departure delays during busy periods. Field demonstrations involving Pacer's previously designed rule-based prediction method showed that the prediction accuracy of departure demand still has room for improvement. This research strives to improve prediction accuracy from two key aspects: better data sources and robust forecasting algorithms. We leveraged two data sources, Aviation System Performance Metrics (ASPM) and System Wide Information Management (SWIM), as our input. We then trained forecasting models with DL techniques of sequence to sequence (seq2seq) and seq2seq with attention. The case study has shown that our seq2seq with attention performs best among four forecasting algorithms tested. In addition, with better data sources, seq2seq with attention can reduce mean squared error (mse) over 60%, compared to the classical autoregressive (AR) forecasting method.

Machine Learning

What problem does this paper attempt to address?

The problem that this paper attempts to solve is to improve the accuracy of aviation flight departure demand prediction. Specifically, the authors applied deep learning (DL) techniques to predict flight departure demands within the strategic time range (4 hours or more) to support Pacer, a mobile application developed by MITRE. Pacer aims to provide general aviation (GA) pilots with future departure demand information, enabling them to better understand potential departure delays. ### Main Problems and Improvement Directions 1. **Improve Prediction Accuracy**: - Previous rule - based prediction methods had insufficient accuracy when predicting departure demands. - Improve prediction accuracy by introducing better data sources and more powerful prediction algorithms. 2. **Optimization of Data Sources**: - Use two main data sources: Aviation System Performance Metrics (ASPM) and System - Wide Information Management (SWIM). - ASPM provides daily - updated historical data, while SWIM provides near - real - time information, thereby reducing prediction errors and increasing prediction frequency. 3. **Adopting Advanced Deep Learning Algorithms**: - Introduce the sequence - to - sequence (seq2seq) model and its variant with the attention mechanism (attention). - These models have been successful in the field of natural language processing (NLP), such as in machine translation tasks. ### Specific Objectives - **Improve the Prediction Model**: Improve the prediction accuracy of aviation departure demands by using the seq2seq and seq2seq with attention models. - **Reduce Prediction Errors**: Compared with the traditional autoregressive (AR) method, these deep - learning models can significantly reduce the mean - squared error (MSE). - **Adapt to Multiple Airport Environments**: Expand the model to other airports through techniques such as transfer learning. ### Experimental Results Through a case study of McCarran International Airport (LAS) in Las Vegas, the authors found that: - **The seq2seq with attention model combining ASPM and SWIM data performed the best**, with its MSE 62% lower than that of the AR model and 39% lower than that of the linear regression (LR) model. - **Even when only using ASPM data, the seq2seq with attention model can still produce the lowest prediction error**. In conclusion, this study demonstrates the superior performance of deep - learning methods in aviation departure demand prediction and provides a basis for further research.

Deep Learning for Flight Demand Forecasting

Flight Demand Forecasting with Transformers

Hybrid Approach Combining Modified Gravity Model and Deep Learning for Short-Term Forecasting of Metro Transit Passenger Flows

Long and Short Term Maneuver Trajectory Prediction of UCAV Based on Deep Learning

Flight Short-Term Booking Demand Forecasting Based on a Long Short-Term Memory Network

Flight Time Prediction for Fuel Loading Decisions with a Deep Learning Approach

Daily Long-Term Traffic Flow Forecasting Based on a Deep Neural Network

Short-Term Forecasting of Passenger Demand under On-Demand Ride Services: A Spatio-Temporal Deep Learning Approach

Predicting the Skies: A Novel Model for Flight-Level Passenger Traffic Forecasting

Research on intelligent forecasts of flight actions based on the implemented bi-LSTM

Trajectory Forecasting with Neural Networks: an Empirical Evaluation and A New Hybrid Model

A novel intelligent approach for flight delay prediction

Online Car-Hailing Supply-Demand Forecast Based on Deep Learning.

A Machine Learning Based Approach for Prediction of Actual Landing Time of Scheduled Flights

A deep learning-based approach for predicting in-flight estimated time of arrival

Flight Trajectory Prediction Based on Hybrid- Recurrent Networks

A hybrid machine learning-based model for predicting flight delay through aviation big data

Phased Flight Trajectory Prediction with Deep Learning

DeepSTCL: A Deep Spatio-temporal ConvLSTM for Travel Demand Prediction

Flight Delay Propagation Prediction Based on Deep Learning

Automated data-driven prediction on aircraft Estimated Time of Arrival