Control and user plane split (CUPS) aided ultra-dense network (UDN) is considered as a promising technology in the fifth generation (5G) networks and will undoubtedly develop in beyond 5G (B5G), which can reduce control overhead and improve the system throughput of network. However, the serious interference and increased power consumption of ultra-dense networking due to densely deployed base stations are also inevitable. To mitigate the interference and improve the system energy efficiency (EE) in CUPS-based B5G UDN, we propose a cluster-based energy-efficient joint user association and resource allocation scheme. First, an improved clustering algorithm is introduced in base station clustering stage. We use orthogonal resource allocation within cluster to reduce co-tier interference and reuse the frequency resource in different clusters. Then, a joint user association, sub-channel allocation and power coordination problem is modeled as a mixed-integer programming problem, which is transformed into a convex optimization problem by using alternative optimization and variable relaxation methods. Finally, we analyze the complexity of our proposed algorithm. Simulation results show that our proposed algorithm can reduce the interference and approximately boost 88% on EE of CUPS-based B5G UDN.

控制和用户平面拆分（CUPS）辅助超密集网络（UDN）被认为是第五代（5G）网络中的一项有前途的技术，并且无疑将在5G（B5G）之外发展，这可以减少控制开销并改善系统网络吞吐量。但是，由于基站密集部署，超密集网络的严重干扰和功耗增加也是不可避免的。为了减轻干扰并提高基于CUPS的B5G UDN中的系统能效（EE），我们提出了一种基于集群的能效联合用户关联和资源分配方案。首先，在基站聚类阶段引入了一种改进的聚类算法。我们在群集内使用正交资源分配来减少共层干扰，并在不同群集中重用频率资源。然后，将联合用户关联，子信道分配和功率协调问题建模为混合整数规划问题，通过使用替代优化和变量松弛方法将其转换为凸优化问题。最后，我们分析了所提出算法的复杂性。仿真结果表明，我们提出的算法可以减少干扰，并在基于CUPS的B5G UDN的EE上提高约88％。