Research Of Dependent Task Offloading And Resource Allocation In Edge Computing

With the development of mobile communication technology and the popularization of smart devices,a large number of emerging applications emerge as the times require,and people have higher and higher requirements for application service quality.Therefore,how to reduce the response time of services and prolong the service life of mobile devices has become a common concern in academia and industry.As an emerging computing paradigm,mobile edge computing can greatly reduce latency and system energy consumption,thereby providing a mobile user experience.One of the challenges of mobile edge computing is to offload computing tasks on mobile devices to the cloud-edge environment under the collaboration of cloud-edge-device and make efficient and reasonable offloading decisions.However,the existing works still have problems such as unreasonable system modeling and ignoring the dependencies between tasks,resulting in that the modeled system model cannot reflect the actual scene,and the resulting unloading decision is infeasible.In response to the above problems,this paper studies the problem of dependent task offloading in the scenario of cloud-edge collaboration.The goal of this study is to find an efficient and reasonable offloading decision to reduce the application’s maximum completion time and system energy consumption.Since this problem is NP-hard,this paper designs and proposes a task offloading algorithm MOIG(Multi-Objective Iterated Greedy Algorithm,MOIG)based on convex optimization and multiobjective optimization,and performs experimental verification in different scenarios.The algorithm proposed in this paper mainly has the following improvements:(1)This paper designs and proposes an algorithm based on convex optimization and multi-objective optimization for the problem of dependent task offloading and resource allocation in the scenario of cloudedge collaboration.The problem studied in this paper is an integer linear programming problem(Integer Linear Programming,ILP)in mathematical form,which is NP-hard.Therefore,by analyzing the constraints and structural characteristics of the original problem,and considering the influence of some constraints on the optimal solution,the original problem is appropriately relaxed,and the convex optimization technique is used to solve it.(2)This paper designs and implements an efficient rounding method,which simply and efficiently restores the optimal solution of the convex optimization problem to the solution of the original problem.(3)For multi-objective optimization problem,this paper designs and implements a heuristic algorithm based on Pareto optimization based on the idea and foundation of genetic algorithm.Finally,based on the above algorithm,this paper designs and implements an edge computing offloading system.This system focuses on the characteristics of edge computing and the needs of mobile applications,provides mobile application developers with low-latency and low-energy consumption edge computing offloading decision-making solutions,and provides network service providers with simple and efficient edge computing resource management services.The system is designed and developed based on Vue.js and Spring Boot,and realizes functions such as user login and registration,information management,computing instance management,computing offloading,and result display,and has complete system functions.In summary,this paper studies the problem of computing offloading of dependent tasks under cloud-edge collaboration,and designs and proposes the MOIG algorithm.When the number of tasks is small,it will increase by ten percentage points,and when the number of tasks is large,it will increase by five percentage points.This paper also designs and implements an edge computing offloading system,which aims to integrate the resources of edge computing and provide a solution for mobile application developers to host applications.After testing,the system has complete system functions,can efficiently manage edge computing resource instances,and takes advantage of edge computing to provide mobile application developers with low-latency and low-energy services,meeting the needs of mobile applications and reaching expected result.