Abstract:The joint resource allocation issue in OFDMA systems or relay-based OFDMA systems with proportional fairness constraints is a mixed discrete programming issue including nonlinear equality constraints, which is hard to tackle efficiently and near optimally. In this paper, an algorithms to solve these two non-convex problems are proposed by Lagrangian dual relaxation based on the equivalent forms. The key in the algorithms is the so-called asymptotic strong duality for OFDMA systems. The asymptotic strong duality suggests that it may be possible to devise polynomial time approximation schemes. These algorithms have the computational complexity linear with the number of subcarriers. Numerical results show that the proposed algorithms can be near optimal to the globally optimal solutions.