Relay nodes (RNs), as important components of heterogeneous networks introduced in Long-Term Evolution-Advanced systems, are employed to improve network capacity and extend cell coverage in a cost-effective manner. In heterogeneous relay networks, there are two types of user access connections: the direct link and the two-hop link. The former is used when a user accesses the base station (BS) directly, whereas the latter, which is composed of a backhaul link and an access link, is used when a user accesses the BS via an RN. For the RNs with in-band wireless backhauls, two-hop links operate in shared spectrum, and thus, resource allocation is a challenging issue to provide sufficient capacity for backhaul links while maintaining fair sharing of resources with normal macro users.

In this paper, we develop a hierarchical game based on the Stackelberg model to address the resource allocation in heterogeneous relay networks. The proposed game consists of two subgames, namely, the backhaul-level game (BLG) and the access-level game (ALG), respectively. In the game model, RNs in backhaul links as leaders play the BLG, whereas mobile stations (MSs) in access links as followers play the ALG. By estimating the achievable rates of MSs, the leaders choose their optimal resource-allocation strategies, and then, the followers do the best responses to the leaders’ strategies. Simulation results demonstrate that our proposed hierarchical game can guarantee the throughput balance between backhaul and access links, and thus, improve the user data rates and the overall system resource utilization, as compared with the resource-allocation schemes that consider backhaul and access links separately.