Coverage and Capacity Optimization in Cellular Radio Networks with Advanced Antennas

We are facing a highly complex optimization problem when using 2-dimensional antenna arrays for concurrent coverage and capacity optimization, because at each single array a multitude of phases needs to be set in order to achieve best performance. Considering an array consisting of 8 x 8 elements and 10 different possible phase values we already have 10(exp 64) possible settings. However, managing this complexity is desirable since the multitude of phase settings available leads to a high degree of freedom when shaping the antenna beam, which in turn, is expected to enhance coverage and capacity performance. We propose a solution technique that reduces the complexity of the problem while still achieving adequate performance gains. We achieve the complexity reduction by considering only a limited set of basic beams which we flexibly combine with each other in order to obtain the overall beam pattern of the actual transmission. We propose a Nelder-Mead and a Q-Learningbased approach in order to adjust the remaining parameters, i. e. vertical and horizontal angles of every basic beam. Based on system level simulations we can conclude that 2-dimensional antenna arrays can significantly outperform linear arrays when carrying out concurrent coverage and capacity optimization.