Multi-leader Multi-follower Game Power Control with Utility Learning for Cooperative Relay Networks over Interference Channels

In this study, the power control problem for multi-user and multi-relay decode-and-forward (DF) cooperative networks over interference channels is investigated. To improve efficiency of power assumption in wireless networks with self-interested nodes, a utility learning based multi-leader and multi-follower stackelberg game-theoretic model is proposed to jointly allocate the power of users and relays, where users and relays are modeled as followers and leaders respectively. In order to motivate the relays to participate in the cooperative transmission, we assume that each relay can get proper profit by introducing the price policy in economics. And the competition among the relays is modeled as a non-cooperative game, where each relay maximizes its utility through dynamically adjusting the price factor. Meanwhile, given specific prices, the users also compete with each other for power resource using non-cooperative game. To reduce the complexity and communication overhead of the power control approach, a distributed utility learning algorithm is proposed to update the game policies of users and relays based on local channel state information(CSI). Simulation results demonstrate that the proposed power control algorithm can improve the energy utilization while guaranteeing the performance of the system.