Rajneesh Singh, Nikita Jain
This thesis aims to forecast aircraft demand in the aerospace and defense industry, specifically aircraft orders and deliveries. Orders are often placed by airline companies with aircraft manufacturers, and then suddenly canceled due to changes in plans. Therefore, at some point during the three-year lead time, the number of orders placed and realized deliveries may be quite different. As a result, orders and deliveries are very difficult to predict and are influenced by many different factors. Among these factors are past trends, macroeconomic indicators as well as aircraft sales measures. These predictor variables were analyzed thoroughly, and then used with time series and multiple regression forecasting methods to develop different forecasts for quarterly and annual orders and deliveries. The relative accuracies of forecasts were measured and compared through the use of Theil’s U statistic. Finally, a linear program was used to aggregate multiple forecasts to develop an optimal combination of all forecasts. In conclusion, the methods employed in this thesis are quite effective and produce a wholesome aggregate forecast with an error that is generally quite low for a forecasting task as challenging as this one.
Aircraft Demand Forecasting