Welcome to the InfoSci Platform
Task Offloading in Cloud-Edge Environments: A Deep-Reinforcement-Learning-Based Solution

Task Offloading in Cloud-Edge Environments: A Deep-Reinforcement-Learning-Based Solution

Suzhen Wang, Yongchen Deng, Zhongbo Hu
Copyright: © 2023 |Volume: 15 |Issue: 1 |Pages: 23
ISSN: 1941-6210|EISSN: 1941-6229|EISBN13: 9781668479261|DOI: 10.4018/IJDCF.332066
Cite Article Cite Article

MLA

Wang, Suzhen, et al. "Task Offloading in Cloud-Edge Environments: A Deep-Reinforcement-Learning-Based Solution." IJDCF vol.15, no.1 2023: pp.1-23. http://doi.org/10.4018/IJDCF.332066

APA

Wang, S., Deng, Y., & Hu, Z. (2023). Task Offloading in Cloud-Edge Environments: A Deep-Reinforcement-Learning-Based Solution. International Journal of Digital Crime and Forensics (IJDCF), 15(1), 1-23. http://doi.org/10.4018/IJDCF.332066

Chicago

Wang, Suzhen, Yongchen Deng, and Zhongbo Hu. "Task Offloading in Cloud-Edge Environments: A Deep-Reinforcement-Learning-Based Solution," International Journal of Digital Crime and Forensics (IJDCF) 15, no.1: 1-23. http://doi.org/10.4018/IJDCF.332066

Export Reference

Mendeley
Favorite Full-Issue Download

Abstract

Cloud computing involves transferring data to remote data centers for processing, which consumes significant network bandwidth and transmission time. Edge computing can effectively address this issue by processing tasks at edge nodes, thereby reducing the amount of data transmitted and enhancing the utilization of network bandwidth. This paper investigates intelligent task offloading under the three-layer architecture of cloud-edge-device to fully exploit the cloud-edge collaboration potential. Specifically, an optimization objective function is constructed by modelling the processing cost of all computing tasks. Additionally, asynchronous advantage actor-critic (A3C) algorithm is proposed under cloud-edge collaboration to solve the optimization problem of minimizing the sum of the weights of task offloading delay and energy consumption. Experimental results indicate that the algorithm can effectively utilize the computing resources of the cloud center, reduce task execution delay and energy consumption, and compare favourably with three existing task offloading methods.