View in EDS

Edge artificial intelligence-based affinity task offloading under resource adjustment in a 5G network.

Saved in:
Bibliographic Details
Title: Edge artificial intelligence-based affinity task offloading under resource adjustment in a 5G network.
Authors: Jin, Wang
Source: Applied Intelligence; May2022, Vol. 52 Issue 7, p8167-8188, 22p
Subject Terms: REINFORCEMENT learning, 5G networks, ARTIFICIAL intelligence, EDGE computing, CLOUD computing, QUALITY of service
Abstract: Artificial intelligence, which can provide cognitive services in mobile edge cloud computing to improve the user's quality of experience, has attracted significant attention. Edge computing can replace local computing to acquire low-latency quality of service in 5th-generation networks. Offloading is one of the key techniques in the 5th-generation network mobile edge cloud computing environment. Big data and computationally intensive applications may suffer from inevitable delays caused by resource homogenization competition and resource switching among different user environments. Affinity, which can help to mitigate resource homogenization competition and reduce switching, is a contextual trait that exists widely between data and between tasks. Due to the mobility of terminals and erratic service requests, deep reinforcement learning has been introduced to learn the offloading strategy for offloading offline affinity tasks in a dynamic environment. The conflict between resource demand fluctuations and fixed resource provisioning yields discontent over the edge resource utilization constraint. In our approach, we develop a long-short term memory and dynamic Bayesian network-based algorithm to predict the periodic throughput and computational capacity data. The distributed unit device in the 5th-generation network is opened up/called back accordingly to relax the edge resource constraint and save energy. We combine deep learning and deep reinforcement learning to provide cognitive services for affinity task offloading and dynamic resource adjustment. A thorough evaluation of the compared algorithms is conducted, and the results show that our approach can help reduce the makespan and costs for big data and computationally intensive applications by 36.3 − 48.2%. [ABSTRACT FROM AUTHOR]
Copyright of Applied Intelligence is the property of Springer Nature and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)
Database: Complementary Index
Description
Abstract:Artificial intelligence, which can provide cognitive services in mobile edge cloud computing to improve the user's quality of experience, has attracted significant attention. Edge computing can replace local computing to acquire low-latency quality of service in 5th-generation networks. Offloading is one of the key techniques in the 5th-generation network mobile edge cloud computing environment. Big data and computationally intensive applications may suffer from inevitable delays caused by resource homogenization competition and resource switching among different user environments. Affinity, which can help to mitigate resource homogenization competition and reduce switching, is a contextual trait that exists widely between data and between tasks. Due to the mobility of terminals and erratic service requests, deep reinforcement learning has been introduced to learn the offloading strategy for offloading offline affinity tasks in a dynamic environment. The conflict between resource demand fluctuations and fixed resource provisioning yields discontent over the edge resource utilization constraint. In our approach, we develop a long-short term memory and dynamic Bayesian network-based algorithm to predict the periodic throughput and computational capacity data. The distributed unit device in the 5th-generation network is opened up/called back accordingly to relax the edge resource constraint and save energy. We combine deep learning and deep reinforcement learning to provide cognitive services for affinity task offloading and dynamic resource adjustment. A thorough evaluation of the compared algorithms is conducted, and the results show that our approach can help reduce the makespan and costs for big data and computationally intensive applications by 36.3 − 48.2%. [ABSTRACT FROM AUTHOR]
ISSN:0924669X
DOI:10.1007/s10489-021-02786-5