The size and complexity of mobile communication networks have increased in the last years making network management a very complicated task. GSM/EDGE Radio Access Network (GERAN) systems are in a mature state now. Thus, non-optimal performance does not come from typical network start-up problems, but, more likely, from the mismatching between traffic, network or propagation models used for network planning, and their real counterparts. Such differences cause network congestion problems both in signalling and data channels. With the aim of maximising the financial benefits on their mature networks, operators do not solve anymore congestion problems by adding new radio resources, as they usually did. Alternatively, two main strategies can be adopted, a) a better assignment of radio resources through a re-planning approach, and/or b) the automatic configuration (optimisation, in a wide sense) of network parameters. Both techniques aim to adapt the network to the actual traffic and propagation conditions. Moreover, a new heterogenous scenario, where several services and Radio Access Technologies (RATs) coexist in the same area, is now common, causing new unbalanced traffic scenarios and congestion problems. In this thesis, several optimisation and modelling methods are proposed to solve congestion problems in data and signalling channels for single- and multi-RAT scenarios.
Buscar en Dimension