Radio Resource Allocation in Wireless OFDM Systems

Abstract

In this thesis, we study radio resource allocation problems in wireless OFDM systems using both centralized optimization and game theoretic approaches. In our work we consider only integer numbers for bit-loading for practical purpose. The subcarrier, bit and power allocation (SBPA) problem in single-cell and multi-cell OFDMA systems which can be formulated as mixed integer nonlinear programming (MINLP) is generally highly complex. We proposed methods to convert MINLP to binary linear programming (BLP) which are linear and much easier to solve. We use the optimal solution as the benchmark for game theoretic approaches. The spectrum sharing game is first studied and the existence of NE is proved under the condition that a player¿s payoff is always positive. Then the SBPA in multi-cell OFDMA systems is formulated as a non-cooperative game and the NE existence is investigated. We propose an algorithm to ensure the convergence of the games and a method to improve system performance with interference avoidance.