Optimizing Mobile Radio Access Network Spectrum Refarming Using Community Detection Algorithms

Abstract

Radio spectrum refarming is a cost-effective way to increase capacity for UMTS/LTE without the need to acquire new spectrum. Therefore, it is an important process for today’s mobile operators. As radio access networks (RAN) became more complex, the process of choosing which radio sites to refarm has become complex and time-consuming if done manually. This paper proposes the use of community detection algorithms for automating the process of RAN spectrum refarming. The problem of RAN spectrum refarming can be modeled as a graph in which one needs to isolate nodes which are densely connected. In RAN spectrum refarming it is crucial to keep the interference between sites and technologies at a minimum. Several community detection algorithms were tested for RAN spectrum refarming. Girvan-Newman algorithm showed most accurate results and was used for implementation of optimization solution. In this paper, the authors provided a benchmark of various network and cluster sizes. Result showed that RAN spectrum refarming can be optimized from one to several weeks to roughly few hours per cluster.