UCB-DQN based Joint Beamforming and Power Control for Software Defined UAV Swarm Network

Abstract

In this paper, we address the problem of deep Q-Network (DQN) based joint power control and beamforming for enabling interference suppression in software defined UAV swarm (SD-UAS) network. We first present the optimization model of joint beamforming and power control (JBPC). Then, to solve this joint optimization problem, we make use of UCB exploration in the learning process of DQN. Simulation results validate that the convergence obtained with the proposed UCB-DQN strategy outperform the DQN learning algorithm, which can to raise the exploration efficiency and sequentially speed up the convergence for the JBPC problem. It is benefit to supress interference and enhance the SD -UAS communication performance.