Joint channel and power allocation in dynamic cognitive small cell networks using asymmetric graphical game

Abstract

Nowadays, cognitive small cell networks (CSCNs) have been widely accepted as a promising and efficient method to deal with exponentially increasing demand of wireless data services. As small cells become smaller and denser, inter-cell interference (ICI) at user equipments (UEs), coming from adjacent base stations (BSs), expanded dramatically and has grown more complicated to be managed. In order to improve cellular throughput in randomly changing small cell networks (SCNs), this article investigates the problem of joint channel and power allocation with dynamic active BSs, using dynamic stochastic game-theoretic learning solutions. It is challenging to address this optimization problem especially when active BSs sets are randomly changing from slot to slot. Dynamic asymmetric graphical game for network throughput maximization will be proposed in this article. It is proved to be an exact potential game (EPG) and the best Nash equilibria of this game correspond to the optimal solutions of this optimization problem. To cope with the difficulties of uncertainty and dynamicity in SCNs, a dynamic joint channel and power selection stochastic learning algorithm (DJCPS-SLA) is put forward, which is analytically proved converging to the Nash equilibria of the accurately formulated game upon changing active BSs sets. The performance of the proposed learning algorithm, DJCPSSLA, far surpasses that of random selection algorithm for joint channel and power allocation. Moreover, the dynamics promote the network throughput to some extent. Finally, the simulation results validate the convergence, annotates a tradeoff between dynamicity, optimality, and robustness.