Multi-device Task Offloading with Time-Constraints for Energy Efficiency in Mobile Cloud Computing.

Kaiyang Liu,Jun Peng,Heng Li,Xiaoyong Zhang,Weirong Liu
DOI: https://doi.org/10.1016/j.future.2016.04.013
IF: 7.307
2016-01-01
Future Generation Computer Systems
Abstract:Nowadays, in order to deal with the increasingly complex applications on mobile devices, mobile cloud offloading techniques have been studied extensively to meet the ever-increasing energy requirements. In this study, an offloading decision method is investigated to minimize the energy consumption of mobile device with an acceptable time delay and communication quality. In general, mobile devices can execute a sequence of tasks in parallel. In the proposed offloading decision method, only parts of the tasks are offloaded for task characteristics to save the energy of multi-devices. The issue of the offloading decision is formulated as an NP-hard 0–1 nonlinear integer programming problem with time deadline and transmission error rate constraints. Through decision-variable relaxation from the integer to the real domain, this problem can be transformed as a continuous convex optimization. Based on Lagrange duality and the Karush–Kuhn–Tucker condition, a solution with coupled terms is derived to determine the priority of tasks for offloading. Then, an iterative decoupling algorithm with high efficiency is proposed to obtain near-optimal offloading decisions for energy saving. Simulation results demonstrate that considerable energy can be saved via the proposed method in various mobile cloud scenarios.
What problem does this paper attempt to address?