Semistatic Cell Differentiation and Integration With Dynamic BBU-RRH Mapping in Cloud Radio Access Network

In this paper, a self-organizing cloud radio access network (C-RAN) is proposed, which dynamically adapt to varying network capacity demands. A load prediction model is considered for provisioning and allocation of base band units (BBUs) and remote radio heads (RRHs). The density of active BBUs and RRHs is scaled based on the concept of cell differentiation and integration (CDI) aiming efficient resource utilisation without sacrificing the overall quality of service (QoS). A CDI algorithm is proposed in which a semi-static CDI and dynamic BBU-RRH mapping for load balancing are performed jointly. Network load balance is formulated as a linear integer-based optimization problem with constraints. The semistatic part of CDI algorithm selects proper BBUs and RRHs for activation/deactivation after a fixed CDI cycle, and the dynamic part performs proper BBU to RRH mapping for network load balancing aiming maximum QoS with minimum possible handovers. A discrete particle swarm optimization (DPSO) is developed as an evolutionary algorithm to solve network load balancing optimization problem. The performance of DPSO is tested based on two problem scenarios and compared to genetic algorithm (GA) and the exhaustive search (ES) algorithm. The DPSO is observed to deliver optimum performance for small-scale networks and near optimum performance for large-scale networks. The DPSO has less complexity and is much faster than GA and ES algorithms. Computational results of a CDI-enabled C-RAN demonstrate significant throughput improvement compared to a fixed C-RAN, i.e., an average throughput increase of 45.53% and 42.102%, and an average blocked users reduction of 23.149%, and 20.903% is experienced for proportional fair and round Robin schedulers, respectively.