论文部分内容阅读
航线旅客流量预测是航空公司航线网络优化的关键技术,传统的预测方法包括回归法、时间序列法等面向旅客订座数进行预测,鲜见考虑航线旅客流量数据较强的随机性和持续增长特性。为了解决上述问题,该研究在回归法的基础上分别基于两种不同的参考期进行预测,并提出一种组合预测模型。该模型的构建分为4个阶段:(1)将传统的订座数预测转换为对客座率的预测,并对客座率数据的一阶累加平滑处理,使得研究目标曲线变得平滑且单调;(2)采用DOW策略的回归法模型预测目标年份的数据;(3)以相邻年度同期拟合曲线的点差值来模拟年度增长量,建立预测模型;(4)针对第2、3阶段两种模型的预测结果,取加权平均值,建立新的组合预测模型。该研究选取某航空公司2011—2015年全年XMNPEK航段客座率数据为依据,预测2016年上半年的客座率数据。对比传统的回归法、时间序列法两种模型的预测结果,平均绝对误差由原来的4.76和4.21缩减到3.77,预测的准确性有明显提高。
Airline passenger flow forecasting is the key technology of airline route network optimization. The traditional forecasting methods include regression method, time series method and so on. It is rare to consider the strong randomness and continuous growth characteristics of route passenger traffic data . In order to solve the above problems, this study predicts two different reference periods based on the regression method and proposes a combined forecasting model. The construction of the model is divided into four stages: (1) The traditional forecasting of the number of reservation is converted into the forecast of the rate of attendance, and the first-order cumulative smoothing of the rate of attendance data is made so that the target curve becomes smooth and monotonous. (2) Predict the data of the target year by the regression method of DOW strategy; (3) Simulate the annual growth with the point spread value of the fitted curve of the adjacent years, and establish the forecasting model; (4) The prediction results of the two models take the weighted average and establish a new combined forecasting model. The study selected the passenger traffic data of XMNPEK segment of an airline from 2011 to 2015 as the basis to forecast the load factor of the first half of 2016. Compared with the traditional regression method and the time series method, the average absolute error is reduced from 4.76 and 4.21 to 3.77, and the accuracy of prediction is obviously improved.