Researches On Radio Resource Management In OFDMA Relay Systems

In future wireless systems, there exist two main contradictions, one is the contradiction between the increasing traffic requirements and the limited radio resources, and the other is the contradiction between the requirement of access everywhere and the complicated wireless environments. Orthogonal Frequency Division Multiple Access (OFDMA) technology and relay technology are effective approaches to relieve the two contradictions above, since OFDMA has high spectrum efficiency and potential to mitigate the problem of frequency-selective fading, moreover, cooperative relaying could increase the coverage of communication systems and enhance the link reliability. Therefore, OFDMA relay systems have been under standardization by the IEEE 802.16j and 3GPP LTE-Advanced task groups.In this thesis, the radio resource management (RRM) problem in OFDMA relay systems are investigated. This thesis first discusses the significance of RRM to future OFDMA relay systems, and then analyzes the differences of RRM between conventional cellular and relay-enhanced cellular networks, based on which the foundation of this thesis is established. The main content of this thesis is to propose novel strategies for the four unsolved aspects of RRM problem in OFDMA relay systems, which are abstracted by reviewing the existing researches.To solve the user fairness problem, this thesis proposes the RRM strategy with proportional fairness. A joint optimization problem is formulated for relay selection, subcarrier assignment and power allocation. Since the formulated primal problem cannot be solved directly, this thesis makes continuous relaxation and proves that the dual gap is zero. And then, the primal problem is decomposed into several subproblems and a master problem by dual decomposition method. The subproblems are solved by Karush-Kuhn-Tucker (KKT) conditions, and the master problem is solved by the subgradient method. Finally, the near optimal solution of the primal problem is obtained.To solve the problem of guaranteeing user satisfactions for multi-service transmission, this thesis proposes the RRM strategy with service differentiation. Different utility functions are adopted to describe user satisfactions, and two typical services which are conversation and background services are considered. The objective of RRM is to maximize the sum utility of background service users, while guaranteeing that the utility of each conversation service user is equal to one. The near optimal solution is obtained by the convex optimization theory. As it is difficult to converge to the optimal solution, a heuristic RRM algorithm is further proposed based on the optimal solution.To solve the problem of stabilizing user queues, this thesis proposes the joint queue and channel aware RRM strategy. Most existing studies assume that sufficient data reside at the buffers of source nodes, while the arrival processes of user data are stochastic in practical systems, and it is significant to stabilize user queues. Therefore, the objective of RRM is to maximize the system throughput while stabilizing user queues. A joint optimization problem is formulated for subcarrier pairing, subcarrier-pair assignment and power allocation. The power allocation and subcarrier-pair assignment problems are solved in the dual domain, while the subcarrier pairing problem is solved using Hungarian method. Finally, the near optimal solution is obtained by an iterative method.To solve the problem of controlling elastic traffic rates, this thesis proposes the cross-layer RRM strategy. The performance of networks can be improved by adjusting traffic rates according to radio resources when the traffics are elastic. However, it is a challenging problem to combine rate control with resource allocation. Thus, this thesis adopts the cross-layer method, where the application, MAC and physical layers are jointly optimized with the objective of maximizing the sum utility of elastic traffics. Using decomposition theory, the cross-layer problem is decomposed into two subproblems, which are the rate control subproblem in the application layer and the resource allocation subproblem in the MAC/physical layer. The two subproblems interact through the"price factor". A two level iterative method is used to obtain the near optimal solution of the primal problem.