Research On Computation Offloading Strategy In Mobile Device Cloud

With the rapid development of the mobile terminal business in the knowledge economy track,the mobile Internet business has made great progress.At the same time,mobile terminal represented by smart phones,tablet computers,and vehicle terminals has won widespread favor from the masses and is closely related to everyone's life.This means that the above technologies have provided mobile applications with strong support.At present,mobile applications has penetrated into every corner of people's lives.People use the Internet as a window to observe the world,through which they can obtain a variety of information and awesome services in life.From the current research situation,mobile terminal equipment has made considerable research results in hardware,but the resources owned by traditional terminals are still very limited.The mobile device cloud refers to a small number of mobile devices providing computing,storage and other functions.The cloud is physically close to the user.The devices in the cloud cooperate with each other to complete the user's request.The mobile device can delegate the task.Giving the adjacent available devices within the same network and calculating by the entrusted devices,thereby completing the tasks that are difficult to complete under the limitation of their own capabilities,because of the small distance between devices,the above operation is more flexible,further more,saves energy consumption and reduces time delay.The tireless research of scholars in related fields proves that in a specific environment,entrusting tasks to other mobile devices can achieve the effect of improving performance in terms of energy and time.Therefore,the proposed and utilized mobile device cloud is valuable.The research content of this paper is a series of strategy research on task segmentation,distribution,scheduling,performance analysis,etc.in the Mobile Device Cloud environment.A feasible calculation and unloading scheme is proposed for the system's time delay and energy consumption performance.And these theories are proved by simulation results.This paper focuses on the problem of task offloading strategy and task offloading performance optimization in mobile cloud computing,and makes full use of the advantages of mobile device cloud communication distance.Firstly,we study the optimization goal of achieving the shortest time delay in the process of computing unloading,and considering In the network communication competition problem,we consider the channel bandwidth allocation mode of OFDMA at the physical layer,combined with the task assignment of the offload target and the transmission rate limitation,an NP problem model based on time delay optimization is established.Then,the simulated annealing algorithm is used to obtain the optimal solution of the problem.More in-depth,we consider that the environment in which the target mobile device is located during computing unloading may be very complicated,the system intelligently generates the need to calculate the unloading strategy,and then combines the method of reinforcement learning with the comprehensive performance of time delay and energy consumption as the objective function.The agent nodes are used to perceive available node information,and the task is divided,scheduled and detected.Finally,based on the system stability and scalability of the entire mobile device cloud computing offloading system,we focus on taking use of Lyapunov stability theory to optimize the stability of queue in computation offloading system.And then,model the target problem,and analyze the Lyapunov methods applied in the art and effect characteristics.