Dynamic Offloading for Multiuser Muti-CAP MEC Networks: A Deep Reinforcement Learning Approach

doi:10.1109/TVT.2021.3058995

Dynamic Offloading for Multiuser Muti-CAP MEC Networks: A Deep Reinforcement Learning Approach

2021-03-01

100

OA

AI

George K. Karagiannidis

A

Arumugam Nallanathan

DOI：10.1109/TVT.2021.3058995

Original

Save

View PDF

Abstract

En 中文

In this paper, we study a multiuser mobile edge computing (MEC) network, where tasks from users can be partially offloaded to multiple computational access points (CAPs). We consider practical cases where task characteristics and computational capability at the CAPs may be time-varying, thus, creating a dynamic offloading problem. To deal with this problem, we first formulate it as a Markov decision process (MDP), and then introduce the state and action spaces. We further design a novel offloading strategy based on the deep Q network (DQN), where the users can dynamically fine-tune the offloading proportion in order to ensure the system performance measured by the latency and energy consumption. Simulation results are finally presented to verify the advantages of the proposed DQN-based offloading strategy over conventional ones.

Keywords:

Task analysis

Energy consumption

Vehicle dynamics

Energy measurement

System performance

Time-varying systems

Time measurement

DQN

dynamic optimization problem

MEC

Dynamic Offloading for Multiuser Muti-CAP MEC Networks: A Deep Reinforcement Learning Approach

Abstract

AI Summary

Journal

Organization