Coverage optimisation is an important self-organising capability that operators would like to have in LTE networks. This paper applies a Reinforcement Learning (RL) based Sparse Sampling algorithm for the self-optimisation of coverage through antenna tilting. This algorithm is better than supervised learning and Q-learning based algorithms as it has the ability to adapt to network environments without prior knowledge, handle large state spaces, perform self-healing and potentially focus on multiple coverage problems.
|Title of host publication
|International Symposium on Wireless Communication Systems
|Number of pages
|Published - 28 Aug 2012