基于对称非负矩阵分解的终端区扇区划分方法

为了合理利用终端区空域有限资源，满足各扇区间管制员负荷均衡为目标，提出了基于对称非负矩阵分解的终端区扇区划分。首先，通过建立终端区网络加权图和量化管制负荷，构建终端区扇区划分模型。其次，为保证相邻性和低方差，对扇区单元层次聚类形成“snake”扇区序列，计算扇区单元之间的相似度值，生成相似度矩阵并进行归一化处理，利用对称非负矩阵正交分解对相似度矩阵进行求解，根据求解出的矩阵判断扇区单元所属区域，实现聚类划分。最后，选取现实终端区空域进行仿真验证，证明了对称非负矩阵分解划分方法的有效性。

Terminal Sector Planning Method Based on Symmetric Nonnegative Matrix Factorization

In order to make rational use of the limited resources of the terminal area and satisfy the goal of balancing the controller load among the sectors, a sector division of the terminal area based on symmetric non-negative matrix decomposition is proposed. First, by establishing the terminal area network weighting graph and quantifying the control load, the terminal area sector division model is constructed. Secondly, in order to ensure the adjacency and low variance, the hierarchical clustering of the sector units forms a "snake" sector sequence, calculates the similarity value between the sector units, generates a similarity matrix and performs normalization processing, and uses symmetric non-uniformity. Negative matrix orthogonal decomposition solves the similarity matrix, and judges the area of the sector unit according to the solved matrix, and realizes the clustering division. Finally, the airspace of the actual terminal area is selected for simulation verification, which proves the effectiveness of the symmetric non-negative matrix factorization method.