Open AccessComputation offloading time optimisation via Q‐learning in opportunistic edge computing
Author(s) -
Yang Guisong,
Hou Ling,
Cheng Hao,
He Xingyu,
He Daojing,
Chan Sammy
Publication year - 2020
Publication title -
iet communications
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.355
H-Index - 62
eISSN - 1751-8636
pISSN - 1751-8628
DOI - 10.1049/iet-com.2020.0765
Subject(s) - computation offloading , computer science , edge computing , server , benchmark (surveying) , distributed computing , node (physics) , computation , task (project management) , upload , mobile edge computing , enhanced data rates for gsm evolution , computer network , algorithm , artificial intelligence , operating system , management , geodesy , structural engineering , engineering , economics , geography
The emergence of computation offloading can meet the real‐time requirements of computing tasks with intensive computing demands. In this study, the authors use opportunistic communication to construct a network framework for opportunistic edge computing (OEC) to perform computation offloading. Specifically, OEC forms a computing resource pool near the edge servers in the edge layer by gathering idle computing resources. Firstly, the state of the system is defined by the attributes of the computing task, the execution location of the computing task and the location of the terminal device in OEC. Then the computation offloading time is calculated and learned by selecting different offloading nodes. Finally, an optimal offloading node selection strategy based on the Q‐learning algorithm is obtained. Extensive simulations show that the proposed strategy consumes the minimum computation offloading time compared with benchmark algorithms in aspects of the amount of uploaded data, the total number of CPU cycles of the task and the number of computing tasks.