Optimal and Efficient Graph-Based Resource Allocation Algorithms for Multiservice Frame-Based OFDMA Networks

This paper addresses the resource allocation problem in Orthogonal Frequency Division Multiple Access (OFDMA)-based wireless networks. The resource allocation problem is posed as an optimization problem with individual user constraints. This formulation provides a special structure that lends to efficient solution of the problem. We develop an optimal algorithm based on standard graph theory and Lagrangian relaxation. Based on the special structure of the problem, the proposed resource allocation algorithm attains the optimal solution at a much lower complexity compared to general-purpose optimization algorithms used by previous OFDMA resource allocation approaches. Moreover, the resource allocation problem solved by the proposed algorithm supports practical features such as discrete modulation set and multiple OFDM symbols per resource allocation decision. Furthermore, by assuming even power allocation across the OFDM subchannels, a suboptimal resource allocation algorithm with lower complexity is developed. The proposed algorithms enable the system designer to control the tradeoffs among system performance, system complexity, and the quality of service (QoS) experienced by the users. Extensive simulations are conducted to evaluate the performance and complexity of the proposed algorithms under different system operating conditions.