Cross-Layer Optimization for Fairness in OFDMA Cellular Networks with Fixed Relays

Fairness and resource allocation in OFDMA cellular networks enhanced with fixed relays using single path relaying scheme and using cooperative relaying scheme are studied in this paper. A cross-layer optimization problem that maximizes the minimum throughput among all users in such kind of network is formulated under path selection and PHY/MAC constraints. A two-level Lagrangian dual decomposition method is used to solve this joint fairness and optimization problem efficiently and optimally. In the first level, a dual problem for primal master problem is divided into a fair throughput control problem in high layer and a weighted sum capacity maximization problem in PHY/MAC layer. The former is linear programming and trivial, while the latter is still complex due to coupled power constraints across tones. In the second level, the PHY/MAC layer problem is solved by further decoupling its dual problem into N per-tone problems, which can be efficiently solved for both single path relaying and cooperative relaying schemes. Subproblems in each level interact through dual variables. The numerical results show that through the cross-layer optimization, the advantages of cooperative relaying are fully exploited and the fairness among all users is strictly satisfied.