Towards Optimal Power Control via Ensembling Deep Neural Networks

TL;DR

This paper introduces ePCNet, an ensemble of deep neural networks trained with unsupervised learning to optimize power control in interference channels, outperforming existing methods in sum rate and computational efficiency.

Contribution

The paper proposes ePCNet, a novel ensemble deep learning approach for power control that leverages unsupervised training to improve performance and reduce complexity.

Findings

ePCNet outperforms state-of-the-art methods by 1.2%-4.6% in sum rate.

ePCNet reduces computational complexity compared to existing solutions.

Unsupervised training enables effective learning without ground truth optimal solutions.

Abstract

A deep neural network (DNN) based power control method is proposed, which aims at solving the non-convex optimization problem of maximizing the sum rate of a multi-user interference channel. Towards this end, we first present PCNet, which is a multi-layer fully connected neural network that is specifically designed for the power control problem. PCNet takes the channel coefficients as input and outputs the transmit power of all users. A key challenge in training a DNN for the power control problem is the lack of ground truth, i.e., the optimal power allocation is unknown. To address this issue, PCNet leverages the unsupervised learning strategy and directly maximizes the sum rate in the training phase. Observing that a single PCNet does not globally outperform the existing solutions, we further propose ePCNet, a network ensemble with multiple PCNets trained independently. Simulation results show that for the standard symmetric multi-user Gaussian interference channel, ePCNet can outperform all state-of-the-art power control methods by 1.2%-4.6% under a variety of system configurations. Furthermore, the performance improvement of ePCNet comes with a reduced computational complexity.