摘要
Cellular networks are now rapidly evolving to renewable heterogeneous network (HetNet) architectures, where a mixture of macro base stations (BSs) and small cell BSs coexist within the cellular network. Particularly, these BSs are partly-powered or self-powered by renewable energy sources (RES). Load awareness has been elevated to a central problem for HetNets, while RES also poses relevant technical challenges. Thus, load balancing for RES powered HetNets is quite challenging. Focusing on this, we study the energy-aware and quality of service (QoS)-aware load balancing for this type of HetNets. With the objective of maximizing network-wide weighted utility of load efficiency in a time duration, where the weight is the corresponding amount of consumed resources, we formulate a user association optimization problem. Based on relaxing, discretization and dual decomposition, an optimal offline algorithm is firstly obtained. Then, distributed and low-complexity suboptimal online algorithms are proposed. Last, given load balancing index and call blocking probability as metrics, numerical results evaluate the performance of the proposed algorithms, which demonstrates the accuracy of the theoretical analysis and provides further insights into the design and operation of the energy-aware load balancing system.