Joint Computing and Radio Resource Allocation in C-RAN Systems Under Imperfect CSI

Abstract

Cloud radio access network (C-RAN) is a promising solution for future wireless network due to easy cooperation among remote radio heads (RRHs) and concentrated management in baseband unit (BBU) pool. Virtual machine(VM) in the BBU pool provides computation ability to different users(UE). However, the advantages of centralized scheduling needs to accurately understand the channel state information(CSI) of the served users, which is hard to get due to quantization error, estimation error, etc. In this paper, we focus on the joint radio and computing resource allocation under imperfect CSI, which is more practical. By jointly optimizing beamforming vector, RRH selection and VM-UE computation resource mapping under the constraint of limited fronthaul and imperfect CSI, we aim to minimize the network power consumption (NPC) in both RRHs and BBU pool. Due to the existence of unknown channel error, the number of SINR constraint is uncertain. We propose a simple two-stage iterative algorithm to solve the problem. Firstly, by fixing RRH selection, we solve the two subproblem: RRH beamforming problem and VM-UE computing resource mapping problem. Then the problem is solved by finding best RRH selection to obtain the minimum NPC. Simulation results suggest that our proposed robust radio and computing resource allocation algorithm can improve the system performance and save energy.