Q-learning based heterogenous network self-optimization for reconfigurable network with CPC assistance

With the irreversible trend of the convergence and cooperation among heterogeneous networks, there emerge some important issues for network evolution. One of them is to reconfigure network elements such as cellular base stations (BSs) or access points (APs) of wireless local area networks (WLANs) according to the real-time network environment, in order to maximize the cooperation gain of different networks. In this paper, we consider cognitive pilot channel (CPC) as an assistant to enable cooperation among heterogeneous networks. Based on the widely used reinforcement learning algorithm, this paper has proposed the heterogeneous network self-optimization algorithm (HNSA) to solve the adaptation problem in reconfigurable systems. In the algorithm, distributed agents perform reinforcement learning, and make decisions cooperatively with the help of CPC in order to reduce the system blocking rate and improve network revenue. Finally our simulation proves the anticipated goal is achieved.