Resource demand-acquisition matching optimization for small cell networks: A distributed learning approach

This paper investigates the problem of self-organizing joint sub-channel allocation and power control for small cell networks, taking into account the real demands of users. Concretely, to improve the efficiency of resource utilization and maximize the number of satisfied users in the network, we formulate the problem as a resource demand-acquisition matching game. In particular, a non-monotonic utility function is utilized to model the users´ bounded demands, with the purpose of relieving blind user competition and improving resource utilization efficiency. Then, the correlated equilibrium (CE) is employed to analyze the outcome of the proposed game. Next, a learning algorithm is designed to achieve CEs in a fully distributed manner. Finally, simulation results validate the effectiveness of the proposed scheme and show its performance gains over existing method in improving the number of satisfied users.