Research On Dynamic Resource Allocation Algorithms On OFDM System

With the development of modem mobile communication, the spectral resource is being scare because of the service's expanding from traditional voice to data, multimedia, and interconnection with Internet. So it has been an urgent task to pursue new wireless technology with higher spectral efficiency.Orthogonal Frequency Division Multiplexing (OFDM) is a promising technique with high spectral efficiency. It divides wireless channel into N orthogonal subchannels, then converts high speed serial bits stream to parallel and transmit through lower speed subchannels. So it can support high speed data service, and has strong capacity of combating with multipath interference. Simultaneously, OFDM can dynamically allocate system resources (such as bit, power, etc.) according to subchannel's fading and multipath delay characteristics, and every subcarrier can carry on self-adaptation modulation. Thus, it has much strong flexibility, and can not only improve the utilization efficiency of system resources, but also meet user's different capacity demands.The methods of dynamic resource allocation are researched in this paper. First, the development, advantages and disadvantages, basic principle of OFDM are given. Second, the idea and superiority of dynamic resource allocation is introduced. Then much more researches are made on the power and bit allocation algorithms for single-user OFDM systems and a simplifying algorithm is presented. Finally, the dynamic resource allocation in multi-user OFDM systems is discussed greatly. Several sub-optimal algorithms are researched. Then, an improved allocation algorithm which has proportion-equity is introduced. The main results of this thesis are: 1．A simplifying algorithm is presented for power and bit allocation in single-user OFDM systems. Simulations and the problem analysis are done. 2．An improved allocation algorithm possessing proportion-equity is introduced because of the disadvantage of existing algorithm. Simulation has done to prove that our algorithm has the performance close to the multi-user waterfilling algorithm which is a joint allocation algorithm and has the lower computational complexity than it. Our algorithm also can be used in actual systems.