DRL based binary computation offloading in wireless powered mobile edge computing

This paper considers the wireless powered mobile edge computing combining wireless power transmission (WPT) and mobile edge computing, where the hybrid access point (HAP) uses multiple radio beams to charge multiple wireless devices (WDs) and WD adopts the binary offloading mode to offload computati...

Full description

Saved in:
Bibliographic Details
Published in:IET communications Vol. 17; no. 15; pp. 1837 - 1849
Main Authors: Shen, Guanqun, Chen, Wenchao, Zhu, Bincheng, Chi, Kaikai, Chen, Xiaolong
Format: Journal Article
Language:English
Published: Stevenage John Wiley & Sons, Inc 01.09.2023
Wiley
Subjects:
Algorithms
Computation offloading
data communication
Decoupling
Edge computing
Energy consumption
Energy efficiency
Internet of Things
Linear programming
Mixed integer
Mobile computing
Nonlinear programming
Optimization
Unmanned aerial vehicles
Wireless power transmission
wireless sensor networks
ISSN:1751-8628, 1751-8636
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:This paper considers the wireless powered mobile edge computing combining wireless power transmission (WPT) and mobile edge computing, where the hybrid access point (HAP) uses multiple radio beams to charge multiple wireless devices (WDs) and WD adopts the binary offloading mode to offload computation workload to HAP via FDMA. It is aimed to maximize the sum computation rate (SCR) of WDs by jointly optimizing the binary offloading decision, transmit power of each radio beam, and WPT duration. Due to the strong coupling between the offloading decision and other optimization variables, the SCR maximization is formulated as a mixed integer nonlinear programming problem. To address this challenging problem, an online DRL‐based decoupling optimization algorithm is proposed. Specifically, the original problem is first split into a top‐problem of optimizing binary offloading decision and a sub‐problem of optimizing transmit powers and WPT duration under given offloading decision. Then a self‐learning DRL framework is designed to output the near‐optimal offloading decisions. Finally, for the sub‐problem, based on the Lagrangian dual theory, an efficient approach to fast obtain the closed‐form expression of the optimal solution is proposed. The simulation results show that the proposed DRL‐based algorithm can achieve the near‐maximal SCR with low computational complexity. This paper considers the wireless powered mobile edge computing combining wireless power transmission (WPT) and mobile edge computing, where the hybrid access point (HAP) uses multiple radio beams to charge multiple wireless devices (WDs) and WD adopts the binary offloading mode to offload computation workload to HAP via FDMA. The sum computation rate (SCR) of WDs is maximized by jointly optimizing the binary offloading decision, transmit power of each radio beam, and WPT duration.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1751-8628
1751-8636
DOI:10.1049/cmu2.12658