Knowledge Graph Construction, Management, and Application in Wireless Networks

Wireless networks generate large volumes of heterogeneous data from network elements, user equipment, and management systems, posing significant challenges for effective network monitoring, fault management, and resource optimization. Traditional rule-based or data-driven approaches often lack unified knowledge representation and reasoning capability, limiting their scalability and interpretability. To address these challenges, this paper proposes a knowledge-graph-based framework for wireless network knowledge construction, management, and application. The proposed framework integrates multi-source network data through ontology-driven modeling and rule-based semantic mapping, enabling structured representation of network entities, events, and their relationships. An event-driven incremental update mechanism is introduced to efficiently maintain the knowledge graph in dynamic network environments without full reconstruction. Furthermore, a lightweight reasoning mechanism is employed to infer implicit network states and support intelligent network management decisions. The framework is designed to balance expressiveness and computational efficiency, making it suitable for large-scale wireless networks. To quantitatively evaluate the effectiveness of the proposed approach, extensive experiments are conducted under different network scales. The experimental results demonstrate that the proposed framework consistently outperforms traditional rule-based methods in terms of fault localization accuracy and resource utilization efficiency, while exhibiting lower query latency and better scalability as the network size increases. The results indicate that the proposed knowledge-graph-based framework provides an effective and scalable solution for intelligent wireless network management, with potential applicability to fault detection, resource optimization, and network security analysis.