Multi-dimensional Resource Optimization Technology Based On Mobile Edge Computing

In recent years,while user demands have exploded,computing requirements for different applications have become greater,and latency requirements have become higher.Although traditional centralized cloud computing has a large amount of computing and storage capabilities,but to deal with a large number of delay-sensitive applications,unified centralized processing will not only cause serious network congestion,and data will transmit between terminals and base stations which cannot guarantee the reliability of the transmission link and will consume a lot of delay as users is far away from cloud base stations.Therefore,mobile edge computing collects and utilizes idle resources at the edge of the network,which not only guarantees a certain processing capacity,but also is closer to users,and meets the requirements of delay-sensitive tasks.This article considers the type of mobile Device-to-Device(D2D)communication within the coverage of the base station.Terminal equipment has certain resources and processing capabilities.In addition to offloading user requests to the base station for processing,some users can also choose to divide task requests into different proportions and offload them to other idle users for collaborative processing.D2D connection may be interrupted intermittently because of user mobility during task transmission,resulting in interrupted task transmission.Therefore,for the interaction time model of a pair of D2D users that perform tasks,the impact of user mobility on the successful completion of the task is studied.Finally,the total energy consumption is taken as the objective function,and the joint optimization of offloading mode,task division,and computing resource allocation are considered.The objective function is divided into a convex optimization problem for computing resource allocation and a non-convex function problem about task division,which are solved using convex optimization theory and a flexible proximal alternating direction method of multipliers(ADMM),respectively.Simulation results confirm the effectiveness of the proposed algorithm and the superiority of the scheme.In addition,this article introduces virtualization technology to virtualize resources of different dimensions on the server and allocate virtual machines to process multiple tasks in parallel.Different from the task offloading,for a task request that has been allocated virtual resources,user movement will cause the corresponding virtual machine to migrate.Therefore,the second research content of this paper focuses on task completion time under different migration strategies.When a mobile user leaves the coverage of the source base station,the virtual machine processing the user’s request can be migrated to another base station for continued execution,or kept on the source host for processing,and the results are transmitted to the user in different ways.With the goal of minimizing the delay cost in a single time slot,we design the best migration strategy and select the best matching server,and pay attention to the transmission rate regulation during the migration process.Finally,it is solved by the Bellman equation of Markov decision.The experimental results show that the proposed strategy improves the experimental performance and the effectiveness of the algorithm.