Spectral-efficient resource allocation for mixed services in OFDMA-based 5G heterogeneous networks

One of the principal thrusts of proposed fifth-generation networks is the provisioning of mixed services or heterogeneous quality of service, which includes delay-sensitive and delay-tolerant services. Guaranteeing the fairness in data rate among end users in such a complex network brings many challenges. In this paper, we consider heterogeneous quality of service and fairness in the problem of spectrally efficient resource allocation in orthogonal frequency-division multiple access (OFDMA)-based fifth-generation heterogeneous networks with femtocells. We aim to maximize the total capacity of all femtocells while satisfying the following: fairness constraint for delay-tolerant users, minimum throughput for delay-sensitive users, cross-tier interference threshold of femtocell heterogeneous networks, and exclusive property of OFDMA subchannels. The formulated problem is combinatorial and nonconvex due to the integer constraint of OFDMA subchannels and the nonaffine equality constraint of the fairness. To make the problem more tractable, we propose to maximize the overall throughput over instantaneous data rate instead of power and subchannel. By doing this, the fairness equality becomes affine on instantaneous data rate. The feasible domain of the problem is redefined with the constraints of subchannels readily included. The transformed problem takes the standard convex form, which is much simpler and is easily solved using the Lagrangian dual method. The proposed algorithm has a low complexity and provides a higher spectral efficiency compared with existing algorithms.