摘要
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。
After analysis and research of the common methods of 4-dimensional （4-D） trajectory prediction in air traffic transport management, a new idea of prediction model based on genetic algorithm （GA） of function mining from history data is proposed. To analyze the features of 4 D trajectory data and overcome shortcomings of common single function mining, a new method of function mining based on frequent function set of gene expression programming （GEP） is proposed to meet the difficult of pro- cessing complex data set. This method made genetic programming of the history data set and mining frequent function set, pre- dicted the trajectory of future from the best function. Finally, the experiment of a real flight data proves the availability and accu racy of this method.
出处
《计算机工程与设计》
CSCD
北大核心
2012年第4期1514-1517,1552,共5页
Computer Engineering and Design
基金
国家自然科学基金民航联合基金重点项目(60736046)
关键词
四维轨迹预测
空中交通管制
基因表达式编程
频繁函数集
函数挖掘
4-D trajectory prediction
air traffic control
gene expression programming
frequent function set
function mining
