Research On Computation Migration Technology Under Mobile Edge Computing

Mobile Edge Computing(MEC)deploys computing resources and storage resources to the edge of wireless network,so as to deal with resource-intensive and delay-sensitive application tasks.It can solve the problems of cloud computing,such as high possession of communication resources,prolonged task processing time and large energy consumption of user terminal equipments,etc.Users can offload tasks to edge servers by computation migration,in order to meet the processing requirements of application tasks with low delay and low energy consumption.With the rapid development and wide application of MEC,it has become a major research direction in this field to design an efficient and high-performance computation migration algorithm that meets the application tasks quality of service requirements.The computation migration technologies have been studied under the scenarios of ultra-dense network,cloud-edge-end three-tier network and mobile network in this paper.And the network resources' allocation,tasks' migration decision and users' mobility are considered and optimized comprehensively.The main work and innovation points are as follows:(1)In the ultra-dense network scenario,a computation migration strategy based on the improved cuckoo search algorithm is proposed,considering the task delay and energy consumption.Firstly,the gain maximization problem of users' task delay and energy consumption is constructed.Then,the multi-attribute decision model is used to select the most suitable migration base stations for users to handle migration tasks,and the optimal allocation of network resources is realized.Finally,because the gain maximization problem is a nonlinear mixed integer problem,the adaptive penalty function method is adopted to construct a fitness function about the migration decision scheme.The function solution under the maximum fitness value can be obtained by the improved cuckoo search algorithm,that is,the optimal migration decision scheme.(2)In the cloud-edge-end three-tier network scenario,a joint optimization strategy of resources allocation and migration decision is designed by adding the cloud center into the MEC computation migration structure.Firstly,the optimization problem model is constructed according to the weighted sum of the relative reduction of task delay and energy consumption.Secondly,the priority of tasks is determined and the decision scheme of initial migration is made according to the task data volume and the task maximum delay constraint.Then,the channel allocation algorithm based on balancing transmission performance and the computing resources competition algorithm based on maximizing resource gain are adopted to allocate channel resources and computing resources for migration tasks respectively.Finally,the migration decision scheme of the optimization problem under Nash equilibrium can be obtained by the iterative comparison method on the basis of game theory.(3)In the mobile network scenario,considering the influence of user mobility on the continuity and quality of computation migration service,a edge service migration strategy based on mobility awareness is proposed.Firstly,according to mobile user's movement trajectory and the switching time of small base station,the task delay model is given in three cases,and the total delay minimization problem is modeled.Secondly,the Lagrange interpolation method is used to model the mobile user's movement trajectory.The user's movement trajectory can be predicted in view of its movement direction that obtained by derivation and speed.Then,considering the cost of switching small base station and virtual machine migration,the optimal communication base stations are assigned to the users.The computing resources are properly allocated by the Lagrange multiplier method combined with the relaxation dual method.Finally,the branch-and-bound algorithm is adopted to solve the virtual machine migration scheme with the minimum total delay of the system.