Active antenna systems for centralized self-optimization of capacity in LTE-A

Network Quality-of-Service (QoS) improvements in terms of capacity can be brought about by optimizing antenna tilts. Active Antenna Systems (AAS) are gaining wide interest recently for their agile ability to electronically adjust the beam in the optimum direction based on operators' QoS interests. In our previous work, we had proposed a novel technique for centralized self-optimization using an iterative modeling and optimization framework called as "surrogate-based optimization". This framework involves: 1-the use of a stochastic model which can be used "as a surrogate" (i.e. to replace) the functional relationships between the network parameters and Key Performance Indicators (KPI), and 2-a subsequent use of a pattern search algorithm as an optimization process, in an iterative manner. In this work, we propose to use the surrogate-based framework for optimization of antenna tilts in Long Term Evolution Advanced (LTE-A) networks and compare optimization results of several capacity based objectives. Results using a flow level network simulator reveals the need to jointly optimize multi-capacity-objectives in a centralized manner apart from demonstrating the superiority of the iterative self-optimization methodology in terms of fast convergence, adaptivity and flexibility.