论文部分内容阅读
通过对空中交通运输管理中目前常用的轨迹预测算法的研究比较和分析,提出了利用遗传算法的从历史数据中进行函数挖掘的思想。针对四维轨迹数据特征的分析和传统的单一函数挖掘的局限性,提出了基于基因表达式编程的频繁函数集挖掘的建模方法。该模型方法通过对历史飞行数据进行遗传算法的操作挖掘出数据集中对应的函数关系集合,用较好的函数模型预测未来航迹。以某一航班雷达数据为训练集做实验,结果表明了应用该方法的准确性和可用性。
Through the comparison and analysis of the trajectory prediction algorithms currently used in air traffic management, this paper proposes the idea of using genetic algorithms to mine functions from historical data. According to the analysis of the characteristics of four-dimensional trajectory data and the limitation of traditional single function mining, a modeling method of frequent function set mining based on gene expression programming is proposed. The model method is based on the genetic algorithm of historical flight data to mine the corresponding set of functional relationships in the dataset and to predict the future trajectory with a better functional model. Experiments with a flight radar data as a training set show the accuracy and usability of the proposed method.