UMTS radio network optimisation.

Radio network planning and optimisation include tasks such as propagation prediction, performance evaluation, and configuration optimisation. Unlike Time Division Multiple Access (TDMA) based systems, the network performance of Third Generation (3G) cellular systems cannot be based on signal predictions only. The maximum achievable path loss in a Wideband Code Division Multiple Access (WCDMA) cell is dependent on the cell load, which creates a complex trade-off between coverage and capacity. Given a user distribution, network performance calculation involves optimisation of transmit powers taking into consideration the location of each user. This mathematical problem is Non-deterministic polynomial time (NP) hard even for small and simplified instances and the solution process is time consuming. This research focuses on methods that allow fast and intuitive performance evaluation for WCDMA networks. In this work, a novel Semi-Analytical Model (SeAM) for user-capacity calculation, based on static snapshot input, is proposed. The model uses a centralised approach that allows the identification of users with the highest impact on link quality. These users can be removed from the system before going through the iterative procedure of solving the NP hard problem. Therefore, this combined power control and user removal algorithm converges to the optimum solution in a lower number of iteration steps compared to a conventional stepwise user removal scheme or a distributed system level simulation. This thesis also documents research findings in the area of network performance analysis using realistic instances, classical propagation models, and varying traffic scenarios. Performance indicators calculated on the basis of analytical methods are compared with results generated by an advanced WCDMA simulator in order to estimate the accuracy and validate assumptions related to analytical modelling of the WCDMA air-interface. Network planners are generally faced with the challenge of controlling network coverage and capacity at minimum infrastructure cost while maintaining user satisfaction. In this thesis, a configuration optimisation algorithm, based on a Simulated Annealing framework, is proposed for automatic planning of Node B locations and antenna configurations. The automatic planning software is implemented using a WCDMA simulator and SeAM, respectively, as core algorithm for network performance evaluation and cost function calculation