基于相似日PSO-SVM的机场流量预测

Prediction of Airport Flow Based on Similar Day and PSO-SVM

机场流量的精准预测是实施航班控制策略、提高航班正常性的重要依据。为准确把握机场流量分布及变化趋势,提出一种基于机场相似日的粒子群优化支持向量机流量预测方法。首先,通过对目标机场在不同天气下的运行情况进行历史数据统计,构建相似度矩阵建立灰色聚类模型对机场相似日进行筛选;其次,采用粒子群优化的支持向量机方法对筛选出的相似日样本进行训练,对机场交通流量进行预测;最后,以广州白云机场的的运行情况为例进行验证。研究结果表明,所提方法较相似日-BP神经网络及传统的PSO-SVM预测方法精度分别提高了1.03%和5.28%,预测精度较高、稳定性较好,可充分反映交通流的动态变化。

Accurate prediction of airport flow is an essential basis for implementing flight control strategy and improving flight normality.To precisely grasp the distribution and trend of airport traffic,a particle swarm optimization support vector machine(PSO-SVM)flow prediction method based on airport similar days is proposed.Firstly,through the historical data statistics of the operation of the target airport under different weather conditions,the similarity matrix was constructed and the grey clustering model was established to screen the similar days of the airport.Secondly,the PSO-SVM method was used to train the selected samples of similar days and predict the airport traffic flow.Finally,the operation of Guangzhou Baiyun Airport was taken as an example to verify this method.The results show that its prediction accuracy is 1.03%higher than similar-day-BP-network and 5.25%higher than traditional PSO-SVM.This mothed can be used to fully reflect the dynamic changes in traffic flow with its accuracy and stability.