Hypergraph-Based Autonomous Networks: Adaptive Resource Management and Dynamic Resource Scheduling

Abstract

As development steadily advances and the capacity for practical applications grows, 5G networks are being deeply expanded in many fields, such as manufacturing, transportation and logistics, and agriculture. The rapid growth in the number of devices in application deployments is causing manual methods of managing resources to become inefficient and network operation costs to increase significantly. Therefore, autonomous networks framework that can adapt to dynamic network environments need to be designed to undertake network resource management. This paper investigates the characteristics of 5G networks, and a hypergraph-based 5G autonomous networks framework is proposed to achieve stable interconnection within the system. First, the network topology is accurately represented by introducing the concept of hypergraph, and an RL algorithm is used to optimize the management of network resources based on the network topology. Then, a BERT model is adopted for resource state awareness from the user satisfaction perspective, and a fuzzy decision based collaborative resource scheduling algorithm is designed to improve service quality. Finally, the key challenges that will still be faced and need to be further solved in the future development of 5G autonomous networks are deeply explored.