Abstract: The long-term forecasting of national-wide civil aviation traffic provides basis for the airspace project planning. A gray combination forecasting model is put forward based on the GM (1, 1) forecast model and least squares principle. With the historical data analysis of the national-wide air traffic from the year 1985 to 2008, the forecasting applicability is examined with three different methods, such as the time series forecasting, regression forecasting, and neural-network forecasting. After analyzing the characteristics of Chinese national-wide civil aviation traffic data and contrasting forecasting result, the gray combined forecasting model produces the best effect among these forecasting models.

Key words: air transportation, air traffic volume, long-term forecasting, grey forecasting, combined forecasting, forecasting model

摘要： 全国民航空中交通量的长期预测是空中交通管理部门制定空域规划方案的重要依据. 本文基于GM(1,1)模型和最小二乘法原理，首次提出了空中交通流量灰组合长期预测模型；同时根据我国民航1985～2008年的飞机起降架次的历史数据，研究了时间序列预测、回归预测以及神经网络预测三种常用预测方法对中国民航空中交通量进行预测的适应性；通过分析我国民航空中交通量数据的特点以及各种预测模型的预测结果表明，本文提出的灰组合预测模型在上述各种预测模型中预测精度最高.

关键词: 航空运输, 空中交通量, 长期预测, 灰色预测, 组合预测, 预测模型