Distributed Resource Allocation for OFDMA System with Half-Duplex Relay using Rateless Code

This paper considers the resource allocation problem in the wireless OFDMA systems with a relay node. We consider half-duplex relay in which the relay node cannot transmit and receive simultaneously on the same frequency. As a result, the use of relay node may enhance or degrade the system throughput depending on the instantaneous channel states between the source and relay, the relay and the destination as well as the source and destination. Hence, it is very important to dynamically adjust the resource (subcarrier) allocated to the relay node so that the relay is used only at the right time according to the instantaneous channel states. Conventional approaches dynamically schedule the usage of the relay node in a centralized manner in which full knowledge of the channel states between any two nodes in the network is required. However, perfect knowledge of the channel states at various nodes are very difficult to obtain. In this paper, we shall propose a distributed resource allocation algorithm on the OFDMA system with half-duplex relay by employing rateless code. Based on the ACK/NAK exchanges between the source, destination and relay, the proposed algorithm iteratively allocates resource to the source and relay, and converge to the close-to-optimal allocation within finite steps. The resource allocation algorithm has low complexity and provable convergence property. Asymptotic throughput performance of the proposed algorithm is derived. We show that the system achieves significant throughput gain compared to the point-to-point baseline system without relay as well as the baseline system with random subcarrier allocation.