Research On Space Demand Forecasting Of Air Cargo

Cargo airlines in China are facing the challenge of more foreign excellent enterprises as well with fierce competition on the aviation market. Domestic cargo airlines should keep pace with development, using revenue management approach to save cost and improve the level of return if they want to remain invincible. Demand forecasting, which has been related to the important decision of future development, transport capacity arrangement and market exploitation made by airlines, to the rational formulation of flight scheduling and aircraft-team planning, is a fundamental part of revenue management. Therefore, the study of flight segment volume forecasting is necessary.This paper mainly studied the method of modeling of air cargo space demand forecasting by GARCH model and support vector machine on the basis of thorough comparison of different forecasting methods. The following researches have been studied:Firstly, the characteristics of air cargo industry in China were analyzed and the systems of air cargo revenue management were introduced. Secondly, the cargo data warehouses were established and the objects of space demand forecasting constructed as well by sorting out and analyzing the dates in a certain airline last few years. Thirdly, several traditional forecasting methods whose characteristics were also analyzed and compared were introduced systematically. Fourthly, theoretical knowledge of ARMA model and GARCH model was introduced particularly. Good results were achieved by establishing GARCH model which were based on forecasting model using ARMA model. Lastly, the theoretical knowledge and mechanism of support vector machine were expounded in detail, by which the regression model of support vector machine was deduced. Forecasting model modeling of support vector machine was mainly discussed. The forecasting model based on support vector machine was established, and compared with other methods. The results showed that support vector machine applied to air space demand forecasting has feasibility and superiority.