A Q-learning-based Downlink Power Control Algorithm for Energy Efficiency in LTE Femtocells

Femtocell is introduced to improve indoor coverage, which is beneficial for both users and operators. But it will also inevitably produce interference management issues in the heterogeneous network which consists of femtocells and macrocells. In this paper, a decentralized Q-learning-based power control strategy is proposed, comparing with homogenous power allocation and smart power control (SPC) algorithm. Simulation results have shown that Q-learning-based power control algorithm can implement the compromise of capacity between macrocells and femtocells, and greatly enhance energy efficiency of the whole network.