Exploiting Cell Similarities in a Radio Access Network to Enhance Explainability for Autonomic Network Management Systems

Modern telecommunications networks are highly complex and require constant real-time autonomic configuration to maintain optimum efficiency. A key technique used in autonomic self-correcting networks is identifying elements in the network that are performing worse than other equivalent elements and applying configurations to the problematic elements which correlate with the better performance on the equivalent elements. Where autonomic re-configuration is carried out on this basis it is important that the autonomous agent (AA) can provide a human interpretable rationale for why it carried out the reconfiguration which in this instance is a reason for why it considers specific elements in the telecommunications network to be equivalent to each other. This paper investigates the utility of Explainable AI techniques to describe and evaluate affinities between elements in the network based on performance data.