Interactive CoMP with User-Centric Clustering Based on Load Balancing in 5G Dense Networks

Considering random networks (RNs) in 5G dense networks, coordinated multi-point (CoMP) is a key to mitigate inter-cell interference (ICI) caused by an irregular topology. Since a RN is dynamically changeable, an interactive CoMP with user-centric clustering scheme based on load balancing (ICLB) is proposed in this paper to make a tradeoff between the fairness and the efficiency. With an interactive method based on a parallel requests-responses mode (a dynamic Gale-Shapley algorithm), ICLB has a fast convergence speed. ICLB is composed of three steps: load-shifting, user equipment (UE) clustering requests and cell responses. In ICLB, we first define the cell load and shift factor to alleviate the load of overloaded cells. Then UE clustering requests are generated in parallel based on received power levels. Finally, edge factor and competitive factor are proposed to help the cells make decisions for clustering responses simultaneously. The simulation results show that in both RNs and other typical topologies, ICLB performs well on load balancing, improves SINR especially for cell edge users and makes a better performance compromise compared with other schemes.