Power Control in massive MIMO Networks Using Transfer Learning with Deep Neural Networks

Power control (PC) plays a crucial role in massive multiple-input-multiple-output (mMIMO) networks. There are several heuristic algorithms, like the weighted mean square error (WMMSE) algorithm, used to optimise the PC. In order these algorithms to perform the power allocation they require high computational power. In this paper, we address this problem through the application of machine learning (ML)-based algorithms as they can produce close to optimal solutions with a very low computational complexity. We propose the use of transfer learning with deep neural networks (TLDNN) under the objective of maximising the sum spectral efficiency (SE). The evaluation results demonstrate that the TLDNN approach outperforms the deep neural network (DNN) based PC and is twice faster than the WMMSE based PC.