On Proportional Fair Scheduling in Multi-Antenna Wireless Mesh Networks--Theoretical Analysis

Proportional fair scheduling (PFS) provides good balance between throughput and fairness via multi-user diversity and game-theoretic equilibrium. Very little analytical work exists on understanding the performance of PFS. Moreover, most researches on PFS are for cellular networks and typically use linear rate model or logarithm rate model to simplify the theoretical analysis of PFS. Since the linear rate model only applies to very small SINR, most researchers prefer the logarithm rate model in their study on PFS. While previous work which is based on the logarithm rate model provides good estimate of the PFS throughput in Rayleigh fading single-antenna cellular networks, they are not valid for multi-antenna wireless mesh networks. In this paper, PFS in multi-antenna wireless mesh networks under Rayleigh fading is discussed. Specifically, we assume that orthogonal frequency-division multiplexing (OFDM) and single-input multi-output (SIMO) techniques are used in the network. In addition, a new mathematical analysis of PFS that applies to both Rayleigh and Rician fading is presented. Simulations are conducted to verify our analytic results on PFS in the proposed mesh network. To the best of our knowledge, this work is the first one investigating the PFS problem in multi-antenna mesh networks.