Research On Task Offloading Algorithm For Mobile Edge Computing

In recent years,the rapid development of wireless communication technology and mobile devices has given rise to emerging applications,such as virus scanning,face recognition,automatic driving and cloud game,etc.However,high energy consumption and high delay sensitivity characteristics of emerging applications put forward severe challenges to user equipments(UEs)with limited computing,storage resources and battery life.Mobile edge computing(MEC)technology provides a solution to this problem.MEC technology deploys MEC servers at the edge of the access network which close to UEs,such as the base station(BS)side,to provide UEs with computing,storage,and network resources required by various applications,which can reduce task execution delay and energy consumption,improve task execution performance and user quality of experience(Qo E)by means of task offloading and caching.In an MEC network that supports task offloading and caching,it is of great significance to design an efficient task offloading strategy considering the demands of tasks and the available resources of MEC server.In this paper,the offloading strategy of MEC network tasks is studied,and the specific contents are as follows:The research background and significance of this thesis is first introduced.Then,the concept,architecture,key technologies and typical application scenarios of MEC are summarized.Finally,the relevant researches on task offloading in the existing literature are analyzed and summarized.In an MEC-enabled network,joint task offloading and computation scheduling problem is studied in this thesis.Comprehensively considering task characteristics,user fairness and available network resources,the joint task offloading and computation scheduling problem is modeled as a maximum task completion delay minimization problem.Since the problem is a mixed integer non-linear programming(MINLP)problem,the traditional algorithm cannot solve it directly.Therefore,the single-user scenario is firstly considered in this thesis,and transformed single-user case problem into a convex optimization problem by introducing a new variable.Theoretical analysis and optimization toolbox are used to solve the problem respectively.Then the problem is extended to multi-user scenario,and a heuristic algorithm is presented to solve the problem.Specifically speaking,the proposed heuristic algorithm includes a task execution mode pre-selection algorithm,a task scheduling algorithm based on parallel mechanism and a task offloading algorithm based on task adjustment mechanism to determine the task offloading and computation scheduling strategy.In an MEC-enabled network scenario,the problem of joint task offloading and information caching is studied in this thesis.Comprehensively considering the task characteristics,user demands and available network resources,the jointly task offloading and auxiliary information caching problem is modeled a system utility maximization problem.Since the modeled problem is difficult to solve directly,this thesis first transforms the original problem into three sub-problems,namely,task offloading sub-problem,auxiliary information caching sub-problem and resource allocation sub-problem.Finally,for the above three sub-problems,greedy algorithm,heuristic algorithm and Lagrangian duality algorithm are used to solve the problems respectively,so as to obtain joint task offloading,auxiliary information caching and resource allocation optimization strategy.