Resource Allocation by Submodular Optimization in Programmable Hierarchical C-RAN

In this paper, we consider a downlink cloud radio access network where a hierarchical structure of virtual controllers is in the center and massive remote radio heads (RRHs) in the coverage area are divided into multiple clusters. Each cluster has one local controller and a high-level controller coordinates control plane decisions among local controllers. Moreover, each local controller is equipped with one server for creating virtual machines (VMs) to execute users’ tasks. Then the output data is transmitted to their respective users through RRHs. This work aims to minimize the average delay including task execution delay and signal transmission delay by joint optimization of task scheduling and resource allocation under power constraint. Specifically, resource allocation consists of VM allocation and RRH assignment. Due to the NP-hardness of the joint optimization problem, we translate it into a matroid constrained submodular function maximization problem and propose a heuristic algorithm to find solutions with 0.5-approximation. Numerical results show that the proposed algorithm can achieve a better performance than two baseline algorithms, i.e., VM prioritized allocation algorithm and RRH prioritized assignment algorithm.