Effective task offloading heuristics for minimizing energy consumption in edge computing

Guichang Yin,Runze Chen,Yi Zhang
DOI: https://doi.org/10.1109/iThings-GreenCom-CPSCom-SmartData-Cybermatics55523.2022.00069
2022-01-01
Abstract:Mobile edge computing (MEC) allows user devices to offload computing tasks to edge servers, which are deployed close to users, for fast execution, short delay and energy saving. As a result, user devices save energy, but edge servers consume tremendous amount of energy. Accordingly, besides regarding the amount of energy consumed by user devices, it is also necessary to take into account that of energy consumed by edge servers. In this paper, we propose effective heuristics for offloading tasks from multi-user devices to multi-core edge servers with the objective of minimizing the total amount of energy consumption under the deadline constraint. The proposed heuristics start from initial task sequences that are generated by the well-known long and short task first rules, respectively. Then, tasks from the initial task sequences are scheduled in a greedy manner so that the considered total amount of energy consumption can be minimized without the violation of the deadline constraint. Simulation experimental results show that the proposed algorithms significantly outperform the compared baseline RoundRobin, indicating the good effectiveness of our heuristics.
What problem does this paper attempt to address?