Automatic equivalent model generation and evolution for small cell networks

Automated Self-optimizing Networks (SON) algorithms have been proposed to address and solve the issues related to optimization in small cell networks. However, automatic optimization approaches require precise knowledge of the deployment environment and users behaviors. This information is generally difficult, expensive to obtain and presents significant computational requirements. In this paper we introduce a method that, based on available measurements, enables the automatic generation of an abstract equivalent model and its adaptation to the environment in which the network is deployed. This model can be a key component to mitigate the computational burden and to speed up the convergence of self-learning and self-evolving coverage optimization algorithms.