基于交通特性的扇区动态管理决策方法研究  

Study on Sector Dynamic Management Decision Method Based on Traffic Characteristics

作　　者：吕晓文 袁立罡[1] LYU Xiao-wen;YUAN Li-gang(Nanjing University of Aeronautics and Astronautics,Nanjing 211000,China)

机构地区：[1]南京航空航天大学,江苏南京211000

出　　处：《航空计算技术》2021年第5期69-73,共5页Aeronautical Computing Technique

基　　金：国家自然科学基金项目资助(61501229);中央高校基本科研业务费专项基金项目资助(NS2019054,NS2020045)。

摘　　要：空域动态管理是均衡飞行需求与管制负荷的有效方法,如何发掘符合运行规律的决策规则以避免经验粗放型管理是其关键内容之一。针对典型繁忙扇区提取表征扇区运行状态与管制复杂性的典型特征,构建可量化计算的区域扇区运行指标体系,采用最大相关最小冗余(mRMR)有监督和谱特征(SPEC)无监督特征选择算法综合加权实现特征的筛选和降维,进一步选取决策树算法提取扇区动态开合决策规则,并设计了基于客观交通特性预测的扇区动态开合科学决策模式。结果证明,模型开合扇结果与实际开合扇操作的符合率达到89%,模型开合扇规则与现场资深管制员管制经验具有较高的一致性。Airspace dynamic management is an effective method to balance flight demand and control load,and how to explore decision rulesinline with the operation rules to avoid extensive management of experience is one of its key contents.Aiming at the typical busy sector,objective features are extracted to represent the sector operation state and control complexity,and a quantifiable regional sector operation index system is constructed.The mRMR(minimal redundancy maximal relevance criterion)supervised feature algorithm and SPEC(spectral feature selection)unsupervised feature selection algorithm are comprehensively weighted to achieve the filtering and dimensionality reduction of key features.Furthermore,the decision tree algorithm is selected to extract the sector dynamic separating and merging decision rules,and the typical pre tacticalsector dynamic separating and merging is designed based on objective traffic characteristic prediction.The results show that the coincidence rate between model results and actual operations of sector separating and merging is 89%,and the model rules of sectorseparating and merging are in good agreement with the experience of senior controllers.

关 键 词：动态扇区规划 特征选择 管制规则 决策树 

分 类 号：V355[航空宇航科学与技术—人机与环境工程]