Research On The Off-loading Strategies In Mobile Edge Computing

With the development on public security and intelligent computing,an amount of intelligent terminal devices have been wide-used(such as drones,portable terminals,etc.).But when some computing-intensive tasks involved,it is a challenge for these terminal devices in terms of computing capacities and battery volumes.At the same time,a novel computing framework: Mobile Edge Computing,is proposed to combine local edge devices and remote cloud services,and all computing tasks are categorized into local tasks and remote tasks.This thesis studies the effects of task offloading under different strategies on energy consumption and time-delay,by proposing two different offloading strategies:1)the task offloading strategy in the situations of end-edges.In a typical end-edge system,and it is assumed that there are computing tasks in each time interval(each task must be successfully executed in this time interval or it will be discarded).At the same time,these tasks may contain private data.In order to balance the time-delay and privacy protection of the whole system,we propose a offloading strategy based on Lyapunov theory.In this strategy,we firstly establish a two-tier computing model,and discuss the computing models of time-delay,energy consumption,channel allocation and battery power change on this model.Then the cycle frequency of CPU and allocating channel resources are discussed.Lastly,the offloading algorithm is implemented and demonstrated in terms of the effectiveness.2)the task offloading strategy in the situations of end-edge-clouds.In an and-edge-cloud system,we firstly propose a novel three-tier computing model and analyze the generation of energy consumption and time-delay in the process of task execution.Then,we design a fitness computing method and propose a task offloading strategy based on Particle Swarm Optimization for energy consumption and time-delay optimization.