Intelligent Cloud-Edge Collaboration for Mixed Continuous-Discrete Resource Allocation in Heterogeneous Networks

Joint channel selection and power control (JCSPC) is important to manage the interference in a heterogeneous network (HetNet), which consists of multiple base station (BS) and user equipment (UE) pairs. The JCSPC problem involves in a mixed continuous-discrete resource allocation and is typically NP-hard. Conventional methods usually obtain a quasi-optimal solution of the JCSPC problem in a centralized manner by assuming that the instantaneous global channel state information (CSI) is available. However, it is demanding to collect the instantaneous global CSI in practical scenarios. In this paper, we develop an intelligent cloud-edge collaboration assisted JCSPC algorithm. With the new algorithm, each BS can independently optimize its JCSPC policy with only local information, meanwhile enhance the sum-rate of the whole HetNet. Simulation results show that the proposed algorithm can achieve comparable and even better average sum-rate performance to the quasi-optimal method with a much lower time complexity.