Resource allocation in OFDMA networks based on interior point methods

This paper studies a joint optimization problem of sub-carrier assignment and power allocation in orthogonal frequency division multiple access (OFDMA) wireless networks. A major challenge in solving the optimization problem is non-convexity caused by the combinatorial nature of sub-carrier assignment problem and/or non-convex objective functions. To address the combinatorial complexity, we formulate the resource allocation problem as an optimization problem with continuous variables. We propose a novel approach based on a penalty function method and an interior point method (PM/IPM) to solve the problem. In specific, using a two-step implementation, the penalty method is applied first to convert the non-convex feasible region to a convex one. Then, the interior point method is deployed to solve the problem which is non-convex only in the objective function. To evaluate the performance of PM/IPM, we apply a genetic algorithm (GA) that achieves near optimal solutions of the problem by iterative searching. Copyright © 2009 John Wiley & Sons, Ltd.