Abstract
Task offloading is the hot issue for current network. The rapid development of intelligent terminals brings new opportunity for offloading. This paper studies the problem of task offloading in a mobile edge computing system, where both the fixed devices with power supply of electricity and mobile terminals are computation resources. Considering tasks in an application are not independent, the task-dependency is introduced to describe the relationship of tasks. Then, the offloading details are analyzed to illustrate energy consumption and completion time during task assignment. Based on task-dependency and offloading process, the problem of task offloading is formulated to minimize the completion time of whole applications under energy and time constraints. Since it is the NP-hard problem, the heuristic offloading algorithm is proposed to assign tasks of whole applications. In the proposed algorithm, based on the task urgency, tasks of whole applications are assigned to chosen devices to achieve the low offloading cost during the initial assignment. To reduce the completion time, the relative remaining cost is introduced to reassign tasks under constraints. Simulation results illustrate that the proposed heuristic algorithm significantly reduces the completion time under system constraints.
Highlights
Computational offloading is the hot issue for current network
Applications are migrated from a mobile terminal which possesses limited computation capacity to a server with powerful computation capacity in the remote cloud [1]
Considering the internal relationship of tasks in an application, the sequence of offloading tasks may affect the results of completion time, offloading cost and energy consumption during the offloading process
Summary
Computational offloading is the hot issue for current network. During the offloading process, applications are migrated from a mobile terminal which possesses limited computation capacity to a server with powerful computation capacity in the remote cloud [1]. Considering the internal relationship of tasks in an application, the sequence of offloading tasks may affect the results of completion time, offloading cost and energy consumption during the offloading process. The system architecture is proposed and the offloading process is analyzed to obtain energy consumption and completion time during task assignment. The problem of task offloading is formulated to minimize the completion time of whole applications under energy and time constraints. Since it is NP-hard problem, heuristic offloading algorithm is proposed to assign tasks of whole applications. We define the system architecture comprising many kinds of computation resources such as mobile terminals, edge hosts and a server in the remote cloud.
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
