Research On Resource Allocation And Scheduling Strategies In OFDM Systems

With the rapid development of wireless communications,the contradiction between the limited wireless resource and the increasing Quality of Service(QoS) requirements of services is more and more urgent.It involves various technologies from different layers of wireless systems to relieve this contradiction.In the physical layer,Orthogonal Frequency Division Multiplexing(OFDM) technology has been approved to be very suitable for wideband wireless systems.In the data link layer,resource allocation and scheduling can be used to improve the frequency efficiency and throughput of communication systems.Due to the distinct advantages of OFDM,it provides flexible freedoms for resource allocation.Thus, resource allocation in OFDM systems has attracted great interests in recent years.This thesis first introduces the application of OFDM in future wireless communication systems and lists several key technologies,based on which the foundation and content of this thesis are established.Then,this thesis introduces the resource allocation in single user,multiuser and multicell OFDM systems,including the general problem descriptions and the resolvements,and lists several hot topics currently.The resource allocation problems in multiuser systems are usually very complicated due to the consideration of users' QoS.The focus of resource allocation lies in finding the suboptimal solutions rather than the optimal ones,which can be realized by methods of step,convex optimization and heuristic.This thesis focuses on resource allocation of multiuser systems,and under the guidance of the above hot topics,the research is deeply investigated.Cross-layer resource allocation is a hot topic recently,so this thesis proposes a crosslayer resource allocation framework,where the application,MAC and physical layers are jointly optimized.A low complicated elitist selection genetic algorithm is developed to solve the proposed problem and compared with other typical algorithms for performance comparison,which demonstrates the effectiveness of heuristic method.This thesis also studies the cross-layer opportunistic scheduling scheme where the MAC and physical layers are jointly optimized.Using the stochastic approximation method,the resource scheduling problem corresponding to average parameters is converted to a resource allocation problem corresponding to instantaneous parameters to be solved in each time-slot.The proposed scheduling schemes achieve good fairness and throughput performance simply.Resource allocation in OFDM systems combined with other advanced technologies is another hot topic.To combine with the cooperative communication,the OFDM-based resource allocation schemes in wireless cooperative cellular systems with and without extra relay stations are considered respectively.Without extra relay stations,to exploit the cooperative diversity and the multiuser diversity simultaneously,a two-slot resource allocation structure is constructed to deal with the users' half-duplex constraints.A low- complexity suboptimal algorithm is developed to solve the problem efficiently,which performs the initial allocation first and then the relay selection.Numerical examples demonstrate the feasibility of step method in solving complicated problems.With extra relay stations,the aim is to maximize the downlink capacity with fairness consideration among users.We make continuous relaxation and solve the dual problem of the relaxed problem using a sub-gradient method.An integer-valued solution is then obtained from the result of the sub-gradient method.Numerical results show the reliability of convex optimization method as well as its application strategy.To combine with the cognitive radio technology, the OFDM-based resource allocation schemes in wireless cognitive radio systems considering multiple cognitive radio users are studied for uplink and downlink respectively.An iterative algorithm is proposed to find the optimal Lagrangian multipliers,deepening the use of convex optimization method.Instantaneous channel status information(CSI) is necessary in the resource allocation and scheduling.To reduce the CSI feedback overhead,three threshold-setting schemes are proposed from the aspects of feedback outage probability,capacity loss and feedback utility. Numerical results demonstrate that through proper threshold design,the feedback overhead can be decreased greatly at the cost of very little capacity loss,leading to the effective application of resource allocation algorithms in the downlink multiuser OFDM systems.This thesis studies the resource allocation and scheduling strategies synthesizingly,from simple OFDM systems to OFDM-based complicated environments,from simple algorithms such as heuristic and step to convex optimization,from analysis under perfect assumptions to implementation considerations.All the researches in the thesis will help comprehension of resource allocation and scheduling in OFDM systems and cast some bright light on the design of the practical systems.