AbstractsComputer Science

The increasing interest in wireless connectivity to the Internet has led to new technologies in cellular networks to provide ubiquitous access to users. One of the promising solutions is to deploy wireless relays, equipped with buffers, in different parts of the cellular networks, to improve both coverage and capacity. In this thesis, the goal is to investigate resource allocation in such networks, considering different challenges, system constraints and users' service requirements. First, based on simple reasoning, analytical investigations and intuitive generalizations, we show that the use of buffers at relays improves both throughput and average end-to-end packet delay. Extensive computer simulations confirm the validity of the presented discussions and the derived results. Subsequently, we propose Channel-, Queue-, and Delay-Aware (CQDA) resource allocation policies, which provide quality of service (QoS) for both delay-sensitive and delay-tolerant users, in a multiuser orthogonal frequency division multiple access (OFDMA) network enhanced with buffering relays. Numerical results demonstrate significant improvements in providing QoS through the proposed resource allocation policies compared with the existing algorithms. Moreover, we introduce a perspective based on which we divide the network area, in a relay-assisted OFDMA system, to smaller cells served by the base station (BS) and the relays. Using convex optimization and dual decomposition, we derive closed form expressions for iterative signaling among the serving nodes to decide about resource allocation. The resulted framework provides insights for designing efficient algorithms for practical systems. Next, we introduce important parameters to be considered in the instantaneous problem formulation for data admission and resource allocation in the relay-assisted OFDMA cellular networks. Taking into account several practical constraints, we propose novel and efficient algorithms for deciding about time slot, subchannel and power allocation in a distributed manner. Numerical results confirm the effectiveness of the proposed parameters and algorithms in reaching the objectives and satisfying the constraints. Finally, we propose a novel scheduling policy, which provides queue stability and is efficient and fair in terms of delay. The proposed policy can be used in the scenarios with shared or independent channels at the BS and relays, leads to less overhead, and facilitates decentralized resource allocation.