Multi-Objective Optimization for UAV-Enabled Wireless Powered IoT Networks: An LSTM-Based Deep Reinforcement Learning Approach

In this letter, we study a multi-objective optimization problem in an unmanned aerial vehicle (UAV)-enabled wireless powered internet of things (IoT) system. Our aim is to maximize the system throughput, maximize the total harvested energy, and minimize the total energy consumption of UAV simultaneously. Since these optimization objectives are in conflict with each other partly, it is computationally quite complex to make the flying decision in uncertain IoT networks. To address this issue, we propose an attentional deep deterministic policy gradient (ATDDPG) algorithm to find joint optimization solution for multiple objectives. Simulation results demonstrate that the proposed ATDDPG outperforms a number of state-of-the-art deep reinforcement learning algorithms.