Key air route segments refer to route segments that play a hub-supporting role in the air transportation network of a specific airspace, and whose traffic anomalies (such as saturation or delay) are likely to trigger chain reactions through segment correlations. Identifying these segments and implementing resource allocation bias and targeted optimization for them can significantly improve network operational efficiency and alleviate delays. Current key segment identification methods only consider the physical connection status between individual air route segments, ignoring the spatiotemporal correlation of traffic changes among segments. To address this issue, this paper proposes a key segment identification method based on a segment-related network. Firstly, redundant search paths are reduced through the pruning of low-impact indicators, and an improved FP-Growth algorithm is proposed to analyze segment correlations. On this basis, a segment-related network is constructed, where segments serve as nodes and correlations between segments serve as edges. This network integrates the changing relationship between segment flows and the physical connections between segments, thereby improving the accuracy of identification. Finally, in the segment-related network of the Guangdong-Hong Kong-Macao Multi-Airport Terminal Area, key segments such as Pingzhou-Conghua, Pingzhou-Gaoyao, and Zhuhai-Nanlang are identified using degree centrality, PageRank, eigenvector centrality, and global centrality analysis methods.

Aiming at the characterized networking and clustering of flight conflicts in high-density airspace under free-flight mode, this paper proposes a network-based conflict resolution method that integrates complex network theory and an improved particle swarm optimization (PSO) algorithm. First, a tunable Mixed Indicator (MI) is formulated by combining the Weighted Cycle Ratio (WCR) and node strength (NS) to accurately identify critical nodes within the flight conflict network. Subsequently, based on the circle structure, a Circle Complexity (CC) metric is proposed as a network structural indicator. Furthermore, the Analytic Hierarchy Process (AHP) is employed to incorporate network efficiency, the number of conflicting nodes, and robustness into a comprehensive performance evaluation index. On this basis, the Leader–Follower Particle Swarm Optimization (LDPSO) algorithm is applied to collaboratively optimize the discrete codes representing heading, speed, and composite resolution strategies, thereby achieving efficient resolution for multi-aircraft conflicts. Simulation experiments demonstrate that the proposed MI significantly outperforms traditional centrality metrics in identifying critical nodes. The LDPSO algorithm proves superior to comparative algorithms, including the standard PSO, Genetic Algorithm (GA), and Atomic Orbital Optimization (AOO), in terms of conflict resolution rate, convergence speed, and solution quality. It can achieve the rapid disintegration of large-scale conflict networks with relatively low computational cost, providing a feasible solution for real-time conflict resolution in dynamic airspace.