A demand forecasting and fleet assignment study in the aviation industry

The global airline industry has a great significance in the world economy and plays a crucial role in the lives of millions of people every year. The aviation industry is more easily affected by economic changes and international developments compared to other sectors due to its intense cooperation with the tourism and industrial industries. Announced in 2020 as a global pandemic, the Coronavirus Disease has left the previous crisis behind and caused a great shock to the aviation industry. The travel restrictions applied on a country basis and the unprecedented decline in passenger demand have caused many airlines to stop their operations and forced some airlines to the bankruptcy level. The industry is recovering rapidly but has not yet returned to 2019 levels. In such a sensitive industry, it is obvious that a potential crisis period and the post-crisis recovery period should be carried with a good planning process. The aim of this thesis is to create a decision support system to assist airline managers in planning airline operations during and after the crisis period. In the proposed decision support system, it is aimed to forecast the number of passengers, which is involved in all steps of airline planning processes and determines the process, and to solve the fleet assignment problem according to the forecasted period. In this regard, Long Short-Term Memory (LSTM) and Box-Jenkins method of SARIMA model were used to forecast the number of passengers. Out-of-sample forecasting was performed with the Box-Jenkins method, which performs better in the data set used. Then, a fleet assignment mathematical model that maximizes the daily airline profit was developed by using the obtained predictions as input. As a result of the modelling, weekly flight-fleet schedules and airline earnings were obtained. The actual fleet assignment results of the flight company during the analysis period and the fleet assignment results obtained from the proposed decision support system are compared in terms of the profit criterion. The analyses revealed that the proposed decision support system is efficient in forecasting the number of passengers and solving the fleet assignment problem.