AI Based Network and Radio Resource Management in 5G HetNets

The demand for pervasive wireless access and high data rate services are expected to grow significantly in the near future. In this context, the deployment of Heterogeneous Networks (HetNets) will enable important capabilities, such as high data rates and traffic offloading, providing dedicated capacity to homes, enterprises, and urban hotspots. Despite HetNet technology will be beneficial for future wireless systems in many ways, the massive cells diffusion has as a consequence an exponential increase of the backhaul traffic that can create congestion and collapse the backhaul network. Virtualization of networks and radio access allows the implementation of complex and efficient decisional processes for radio and network resource optimization, but the interaction between lower and upper layers during resource allocation decisions is still mostly unexplored. In this paper we propose an artificial intelligence based approach, with two interdependent decisional cores exchanging information, one aware of physical layer aspects and the other controlling pure network resources. The two iterative procedures aim at jointly optimizing the distribution of the traffic in the backhaul network and the users cell association, with the goals of minimizing the unsatisfied users data rate requests and minimizing the energy consumption reducing the number of activated cells, respectively.