Dynamic Association for Load Balancing and Interference Avoidance in Multi-cell Networks

In promising OFDMA systems, downlink signals originating from the same base station (BS) are orthogonal, while those from different BSs interfere with each other. As a consequence, inter-cell interference (ICI) becomes major performance degradation factor. Particularly, boundary users suffer from severe ICI in addition to the inherent near-far problem. To improve cell edge performances and support a more balanced data rate among all users, partial frequency reuse (PFR) and load-balancing schemes are investigated in this paper. We have formulated a utility maximization problem with network-wide proportional fairness (PF) as an objective in a multi-cell network with PFR. To solve this problem, we propose an offline optimal algorithm and also efficient online algorithms. Our online algorithms are based on simple inter/intra-handover and cell-site selection in which a metric is changed from the signal strength to the average throughput. Through extensive simulations, we demonstrate that our online algorithms can achieve network-wide PF very closely. Compared to the conventional system with a universal frequency reuse where a user is bound to the best signal strength base station, the proposed algorithms bring two types of performance gain: interference avoidance (IA) and load balancing (LB) gain. These gains improve the system performance, especially for users at the cell boundary.