Load balance for centralized heterogeneous networks with limited fronthaul capacity

Load balance (LB) is important for heterogeneous networks (HetNets) composed by low power small cells and high power macrocells. In centralized HetNets (CHetNets), the limit on fronthaul capacity has an innegligible impact on LB performance. This paper focuses on the fronthaul capacity aware LB in CHetNets. First of all, a user cell association optimization problem is formulated to maximize the total rate of CHetNets with the constraint on fronthaul capacity. Then, an improved hybrid genetic algorithm and simulated annealing algorithm (IGASA) is proposed to provide a sub-optimal solution to the problem with low complexity. It is verified by simulations that compared to conventional LB schemes, the proposed fronthaul capacity aware LB scheme with IGASA can improve the performance of CHetNets with various fronthaul capacities, i.e., the outage probability can be reduced as high as 53%, and the user average rate can be increased by up to 51%. Moreover, IGASA can provide good performance close to that of the well-known Sequential Fixing (SF) algorithm with a much lower complexity. Hence, IGASA is more suitable for CHetNets with densely deployed cells.