Deep Reinforcement Learning for Task Offloading and Power Allocation in UAV-Assisted MEC System

Abstract

Mobile edge computing (MEC) can provide computing services for mobile users (MUs) by offloading computing tasks to edge clouds through wireless access networks. Unmanned aerial vehicles (UAVs) are deployed as supplementary edge clouds to provide effective MEC services for MUs with poor wireless communication condition. In this paper, a joint task offloading and power allocation (TOPA) optimization problem is investigated in UAV-assisted MEC system. Since the joint TOPA problem has a strong non-convex characteristic, a method based on deep reinforcement learning is proposed. Specifically, the joint TOPA problem is modeled as Markov decision process. Then, considering the large state space and continuous action space, a twin delayed deep deterministic policy gradient algorithm is proposed. Simulation results show that the proposed scheme has lower smoothing training cost than other optimization methods.

Multi-Agent Reinforcement Learning-Based Resource Management for V2X Communication

“…To address the challenges caused by obtaining the global channel status information, we adopt the reinforcement learning (RL) method (Zhao et al, 2021;Zhao et al, 2020) in this paper. Currently one of the most powerful machine learning tools, RL is usually applied to time-varying dynamic systems (Wu et al, 2018;Yan et al, 2018) and the wireless network (Simsek et al, 2018;Zhao et al, 2018).…”