A comparison of different optimisation search methodologies for self-optimisation in wireless cellular networks

Abstract

Self-Organising Networks (SON) concept is seen as a way to reduce costs by automating functionalities such as network optimisation usually performed manually with extensive human work time. This paper provides a general formulation of the self-optimisation problem in a cellular wireless network and a description of the optimisation search process by means of iterative algorithms. Different optimisation methodologies, namely simulated annealing, genetic and particle swarm algorithms have been considered. These methodologies have been implemented for the optimisation of the cell coverage and cell overlap using real measurements of a UMTS network deployed in a medium-size European city. These methodologies have been compared in terms of speed of convergence and performance of the solutions provided by the different proposed algorithms.