Network Slicing deployment Algorithm for Power Data Network

Studying the network resource slicing mechanism for power data networks, the research addresses the high requirements for reliability, stability, and security in power data network operations. By leveraging Software-Defined Networking and Network Function Virtualization technologies, a physical network is divided into multiple logically segregated virtual networks. A network slicing mapping algorithm is proposed that combines Graph Neural Networks with Deep Reinforcement Learning. This algorithm aims to allocate resources effectively within the network by providing feasible strategies, thereby optimizing the slice creation process for maximum benefits. Experimental results demonstrate that the proposed GNN-A2C model, compared to existing models, enhances resource utilization and revenue ratio, exhibiting stronger generalizability and practical value. These outcomes validate the feasibility and effectiveness of the model.