Robust distributed resource allocation in OFDMA femtocell networks

In this paper, we investigate joint power and subchannel allocation in orthogonal frequency division multiple access (OFDMA)-based femtocell networks under uncertain channel conditions. We present the problem as a Mixed Integer NonLinear Program (MINLP) with quality-of-service (QoS) constraints to provide minimum data rate requirements for femtocell users and avoid co- and cross-tier interference. Previous works on this subject assume perfect knowledge of Channel State Information (CSI) which is unrealistic in practical systems. In this paper, we formulate a robust resource allocation optimization problem in which the co-tier and cross-tier channel gain uncertainties are taken into consideration. In our problem, uncertainty is studied in the form of bounded sets. The effect of uncertainty on the objective function and constraints is demonstrated and the probability bound of constraint violation is derived. Moreover, we propose a fully decentralized robust algorithm utilizing dual decomposition and primal dual update method. We also introduce parameters to address the trade-off between robustness and performance. Our extensive experimental results examine the cost of robustness, show the effectiveness of our robust solution in uncertain environments, and demonstrate the convergence of our proposed algorithm.