Study On Adaptive Resource Allocation Algorithm In OFDMA Systems

As the continuous development of society, people’s demands for wireless communicationservices are not only the voice service, but also the web service, handphone video and onlinetelevision, and they have been an essential part in some subscribers’ daily life. The highertransmission rate is as a consequence required, so that a seriers of new technologies, such asmultiple input multiple output (MIMO), orthogonal frequency division multiple access (OFDMA)and relay have been focused by researchers. However, due to the shortage of radio spectralresources, dynamic resource allocation has also attracted people’s attention.By ultilizing MIMO technology, we can achieve high capacity, improve transmission reliabilityand reduce bit error ratio (BER) in the system. OFDMA is also used to increase systemperformances and combat frequency selective fading. In order to reach1Gbps peak rate fordownlink and500Mbps for uplink, Long term evolution-Advanced (LTE-A) proposes relaycooperatative transmission as one of the key technologies for its high capacity and large coverageof BS. By the way, to promote system capacity and spectral efficiency, we can also use a rationaldynamic resource allocation algorithm.Nowadays, resource allocation based on a lot of practical networks (such as wireless Meshnetworks, Ad hoc networks, Single/Multiple cells communication networks and Realy cooperativenetworks) has been well studied. This is a constrainted optimization problem, including powerminimization subject to rate constraint and rate maximization subject to power constraint. Somebasic algorithms such as water-filling algorithm, greedy algorithm and several intelligent bionicalgorithms are used to solve this problem. As the deeper of research, we should consider thepractical complex networks, so that the algorithm become much more complicated, manyoptimization methods are introduced to this area. Meanwhile, the priority, fairness and proportionalfairness should be taken into account when guaranteeing the rate and power.In this thesis, I study dynamic resource allocation algorithm for three different networks:Downlink single cell MIMO-OFDMA system. A novel linear complexity suboptimal resourceallocation algorithm is proposed based on the assumption that the channel knowledge is perfectlyknown at the BS, and QoS requirement is considered for different services. Downlink single cell multi-relay based OFDMA system. Assume the channel knowledge isknown perfectly at the BS, I propose a proportion fairness optimization model and a new subcarrierand relay selection scheme, and derive the closed-form expressions for power allocation.Due to considering error of channel estimation algorithm and feedback delay, a proportionfairness resource allocation algorithm based on imperfect channel state information (CSI) isproposed in this thesis, and a low complexity joint power and subcarrier allocation algorithm isalso proposed.The three algorithms for different networks are all based on convex optimization theory, andthey have optimal solution or suboptimal solution in theory. Some performance parameters such assystem capacity and proportional fairness are simulated, and the simulation results show that theproposed algorithms perform better than the previous algorithms for the corresponding networks.