Research On Service Selection In Mobile Edge Computing Environment

With the advent of the Internet of Everything and the birth of the fifth-generation mobile communication system(5G),which requires high-bandwidth,high-density,and low-latency service environments,a single cloud computing model may no longer meet the requirements,so the concept of mobile edge computing(MEC)was born as needed.MEC effectively integrates wireless network and Internet technology,and provides a network architecture that can provide users with required services and remote computing,storage and other functions on the wireless side,allowing users to enjoy high-quality uninterrupted service experience.However,MEC also has certain restrictions in the process of providing services to users.On the one hand,because the user's mobility and the signal coverage of the edge server are limited,they can only provide services to users within their signal range.The signal coverage of adjacent edge servers may overlap.During task processing,mobile users may have multiple server choices.On the other hand,the computing,storage,and functional resources configured on the MEC server are limited.When multiple users make simultaneous requests,task overload may occur.The traditional edge server selection problem is often regarded as a static optimization problem.Considering the above situation,how to choose the right and optimal edge server for mobile users is the main content of this paper.The main work of this paper is as follows:(1)Considering the user's mobility,a predictive user path matching server scheme is proposed.This scheme is based on Lagrangian interpolation to fit the user's moving path,and selects the candidate MEC server in the moving process according to the slope of the trajectory polynomial and the distance between the transmission nodes.(2)Considering the user's moving path,a scheme of dynamically selecting the server is proposed.First,establish a model for selecting a server for mobile edge computing.Then,based on this model,the problem is transformed into an optimization problem that minimizes network delay under the constraints of energy constraints.Finally,the edge server with the smallest delay is obtained through the genetic simulated annealing algorithm,which guarantees Smooth user experience.(3)When considering multi-user selection of MEC server,because the resources configured in the MEC server are subject to certain restrictions,in order to prevent task overload,some services can be selected to migrate to the adjacent MEC server.This problem is modeled and transformed into the optimization problem of minimizing system cost.Then an adaptive inertial weight binary particle swarm optimization algorithm is proposed to obtain the optimal unloading decision.Finally,the improvement of the algorithm performance is proved through experiments.