Soft frequency reuse in large networks with irregular cell pattern: How much gain to expect?

Soft frequency reuse (SFR) has been proposed for enhancing cell-edge performance in OFDMA networks. Previous studies of SFR have used simulations over rather small networks having hexagon-shaped cells. In this paper, we examine the expected performance gain of SFR for networks in realistic radio environments and with irregular cell patterns. We define a performance metric that allows for fast assessment of expected gain of SFR over the service area of large networks. An optimization algorithm is used to tackle the resulting frequency allocation problem. Our approach enables SFR to go beyond the conventional schemes in terms of the number of sub-bands and power setting. We provide a thorough numerical study of SFR in large scale networks and investigate the performance trade-off between cell-edge and cell-center zones to illustrate pros and cons of SFR.