Research On Computation Offloading And Resource Allocation Optimization For Mobile Edge Computing

Mobile edge computing(MEC)is an innovative computing paradigm which pushes rich computing and storage resources to the edge of networks.MEC can make up for the shortage of cloud computing and provide a new solution to overcome problems,e.g.,device with limited computing and storage resources cannot meet the computation requirements of emerging applications.Moreover,it has attracted great attention from academic and industry researchers.Devices transfer the computation-intensive mobile applications to the edge servers by offloading to meet the computing and delay requirements of computation tasks effectively.Therefore,in order to make full use of limited computation and communication resources for meeting more computation requirements of users and improving the quality of service,it is necessary to coordinate the computation offloading among multiple users in MEC offloading,which brings a series of new challenges and higher requirements for designing accurately and effectively computation offloading schemes.In addition,among the available affluent computation-intensive applications,the limited computing resources of MEC servers cannot fully attain all computation offloading requirements,and it becomes increasingly prominent with the growth of the number of computation tasks.Therefore,designing a proper collaborative computation offloading scheme to fully coordinate the MEC and cloud computing resources and satisfy the service requirements of users is of significant interest.To solve the above problems,this thesis focuses on the issues of computation offloading and resource allocation in MEC.Firstly,starting with the resource management scheme,this paper reveals the importance of joint optimization of computing offload decision and resource allocation,and designs a computing offload and resource allocation optimization scheme for multi-access point and multi-user MEC system.Next,partial offloading in spectrum sharing of two-layer heterogeneous networks is deeply studied by taking the idea of task division into consideration.Then,three-layer collaborative computation offloading and resource allocation schemes in heterogeneous networks are studied to explore performance gains brought by collaboration between MEC and cloud computing.Finally,the research on computation offloading and resource allocation in the mobile network scenario is further extended to that for the vehicular networks(VANET).The main research results are listed as follows:1)Computation offloading and resource allocation optimization for multi-access point and multi-user MEC system.There are many problems in multi-access point multiuser MEC system,e.g.,low offloading efficiency and high cost,which are caused by unbalanced load and unreasonable allocation of limited computing and communication resources.Therefore,to solve the problem of resource allocation among multi-user,reasonable resource allocation is proposed under the constraints of computing resources,transmission power,and bandwidth.Moreover,we model access point selection as an offloading decision,analyze the unbalance load problem utilizing computation offloading strategy,and further formulate a cost minimization problem.In order to solve the proposed optimization problem,a computation offloading and resource allocation scheme is proposed,and the optimal offloading strategy and the rational allocation of computation and communication resource are further obtained.Simulation results show that the proposed scheme realizes the load balance among access points and achieves the lower system cost.2)Computation offloading of spectrum sharing in the two-tier heterogeneous networks is then explored.Heterogeneous networks can effectively improve data transmission rate,and partial computation offloading can significantly reduce the task completion time.For this reason,a partial computation offloading model for spectrum sharing in the two-tier heterogeneous networks is proposed.By analyzing the parallel processing time of computation tasks in partial computation offloading,the optimal offloading ratio can be obtained,and then a rate optimization problem is presented.Aiming at the strict interference constraints in optimization problems,a priority-based power allocation scheme and an iterative search-based power allocation scheme are designed to optimize the power control.The proposed schemes can improve the data transmission rate and reduce the task completion time while satisfying the interference constraints of the MBS.3)Three-tier computation offloading and resource allocation optimization in heterogeneous networks.Poor scalability and limited computing resources are key issues that need to be solved when MEC provides offloading services for a large number of users.To this end,a three layers computation offloading framework,including users,MEC and cloud computing,is designed for heterogeneous networks to achieve the cooperation of cloud computing and MEC,and provide wide computation offloading service for users.Then,based on the proposed framework,this problem is formulated as a constrained optimization problem for minimizing system cost.To solve this optimization problem,a cost-efficient computation offloading and resource allocation optimization scheme is designed,which jointly optimizes transmission power and computation resource allocation.We first transform it into a quadratic constrained quadratic programming problem,and then obtain the offloading strategy by combining positive semidefinite relaxation and randomization.Experimental results show that the proposed scheme can not only meet the computing needs of users better,but also significantly improve the performance of system.4)Computation offloading and resource allocation optimization for cloud assisted MEC in VANET.As a typical application scenario of MEC,the deployment of MEC in the vehicular networks can effectively improve the limited data processing and storage capacity of vehicles,and provide a guarantee for applications with low delay requirements such as automatic driving of vehicles.To this end,computation offloading and resource allocation for the coexistence of MEC and cloud in vehicular networks is studied.Considering stricter delay constraints and vehicle position changes during offloading,a utility function,including task completion time,computing resource costs and standardization factors,is designed to measure the satisfaction of vehicles for completing computing tasks under different computing offloading decisions.Then,a system utility maximization problem is proposed.A collaborative computation offloading and resource allocation optimization scheme is designed to achieves the optimal solution.Simulation results show that the proposed scheme can effectively improve the system utility and decrease task completion time.