Enhancing HSDPA Performance Via Automated and Large-Scale Optimization of Radio Base Station Antenna Configuration

The rapid deployment of High Speed Downlink Packet Access (HSDPA) calls for automated optimization in network planning. In this paper, we study HSDPA performance improvement by means of optimizing base station antenna configuration. We consider networks with mixed HSDPA and Release 99 (R99) services that share the power resource of the cells. We present an optimization framework to capture the relationship between antenna configuration, service coverage, power sharing between R99 and HSDPA, and HSDPA performance, taking into account the influence of cell size and user distribution on HSDPA user throughput. The cost function is targeted at improving the average HSDPA user throughput. The engine of our computational machinery is a simulated annealing algorithm that is able to search and improve antenna configurations effectively and time-efficiently. We report the benefit of our approach for a realistic planning scenario for the city of Lisbon. The experiment demonstrates that automated optimization leads to significantly better HSDPA throughput.