The Study On Dynamic Resource Allocation Algorithms In OFDM Systems

With the fast development of science and technology, the requirements of the service quality and kinds of business are becoming much higher, therefore, future wireless mobile communication technology shall seek progress from many aspects. Orthogonal Frequency Division Multiplexing(OFDM) is a widely used technique for the next generation wireless communication systems, due to its strong ability to against interference, high frequency efficiency and suitable for high speed transmission. In OFDM systems, the broadband wireless channel is divided into some orthogonal narrow band sub-channels, and each sub-channel suffers different channel fading with different channel capacity. Therefore, in order to achieve higher frequency efficiency and better system performance in limited bandwidth, we must find efficient wireless resource allocation schemes to allocate different bits and power to different sub-channels according to the channel states.In this thesis, the research background, principle and two optimal criteria of dynamic resource allocation are presented at the beginning. Based on these information, we discuss the broadcast characters of wireless channels, introduce the basic theory, system implementation, the advantages and disadvantages of OFDM systems, and list some key technologies. Then the system structure of single-user OFDM system is analyzed, and current research results are listed. Optimal single-user allocation algorithms like Greedy, Campello’s, Piazzo’s and Krongold’s, and one sub-optimal Jang’s algorithm are presented with basic concept, flow chart and advantages and disadvantages. Then the structure of multi-user OFDM system is given, followed by the state of the art. The optimal algorithm and sub-optimal algorithms like Wong’s, Max-min and Zhao’s are studied. Based on the abovementioned analysis, two improved algorithms, the algorithm using the dynamic boundary constraint method and the improved algorithm based on Zhao’s conversion algorithm, are proposed. The two improved multi-user algorithms are analyzed and simulation results are given. Simulation results show that both algorithms perform better than that in the literature in proportional fairness, complexity and system capacity.