Abstract:
The focus of this thesis is on end-to-end (e2e) queueing performance evaluation and resource allocation in order to improve the performance of the relay-enhanced cellular networks. It is crucial to study both the performance of the data link layer and the physical layer issues. Therefore, we first consider end-to-end queueing performance evaluation and after that to consider physical layer issues, we present power allocation schemes, relay load balancing and relay assignment. First, we presented a framework for the link-level end-to-end queueing performance evaluation. Our system model consists of a base station, a relay, and multiple users. The e2e system is modeled as a probabilistic tandem of two finite queues. Using the decomposed model, radio link-level performance measures such as e2e packet loss rate, e2e delay and throughput are obtained analytically and compared with simulation results. A framework for power allocation for downlink transmissions in decode-and-forward relay networks is investigated. We consider a system with a single base station communicating with multiple users assisted by multiple relays. The relays have limited power which must be divided among the users they support in order to maximize the data rate of the whole network. Based on knapsack problem, the optimal power allocation is proposed. To consider fairness, weighted-based scheme is presented. Moreover, to utilize the power wisely, an efficient power reallocation scheme is proposed. Simulation results demonstrate the efficacy of the proposed schemes. By applying the relay selection scheme, it may happen that some relays have more users connected to them than other relays, which results in having unbalanced load among the relays. In order to address this issue, a game theoretic approach is presented. Coalition formation game is proposed based on merge-and-split rule to form the optimal structure. The simulation results demonstrate the effect of applying game in proposed problem. Finally, the relay assignment procedure is studied. The optimal solution is found using Lagrangian Relaxation. Then, a lighter algorithm is proposed to efficiently carry out the relay assignment. Simulation results show that the proposed algorithm can achieve near optimal data rate, while it decreases the processing time significantly.
