Research On Offloading Algorithms Of Mobile Edge Computing In Heterogeneous Networks

With the rapid development of mobile services,user equipment with limited computing,storage and battery resources has been unable to meet the demands of high-complexity and high-energy-consuming services.Mobile Edge Computing(MEC)enables user equipment to offload tasks to edge servers by migrating computing,storage and service functions to the edge of the network,thereby reducing the completion delay and energy consumption of user equipment tasks.How to make efficient offloading decisions and resource allocation to reduce task completion delay and energy consumption has become an important research topic.In this thesis,user task offloading and resource allocation in heterogeneous cellular networks are studied,and a joint optimization algorithm for base station selection,computing offloading and resource allocation in multi-base cellular networks is proposed.Combining Device to Device(D2D)communication with MEC,a multi-user task offloading algorithm based on graph theory is proposed.Based on wireless power supply technology,a multi-user offloading decision and task scheduling algorithm is proposed.The main research contents and innovations of this thesis are as follows:(1)For the computing and offloading scenario composed of users,edge servers and remote cloud servers,a joint optimization algorithm of base station selection,computing and offloading and resource allocation in multi-base station cellular network is proposed.This algorithm considers the base station selection of multiple base stations overlapping users,and minimizes the weighted sum of energy consumption and time delay under the constraint of edge server computing resources.Firstly,the Lagrange multiplier method is used to solve the offloading problem of single user and multiple base stations.Based on base station scene,many users consider user base station selection and edge server computing resources competition,access to the base station are chosen based on the definition of the choice function,adopt suboptimal iterative heuristic algorithm for single user scenarios of offloading decisions dynamic correction,uninstall decision-making and edge server resource allocation.Simulation results show that the proposed algorithm can effectively reduce the delay of task completion and terminal energy consumption.(2)For MEC scenarios assisted by D2 D communication,a multi-user task offloading algorithm based on graph theory is proposed.According to user tasks completed by the constraints of time and energy,to delay sensitive user computing tasks into tasks and sensitive task of energy consumption,based on graph theory to idle equipment,the edge server,and remote cloud server,respectively,to build the corresponding delay weight and energy consumption weight figure,and defines the optimization goal for time delay and energy consumption.The maximum matching and minimum cost graph algorithm is used to solve the offloading strategy.Firstly,the delay weight graph is solved to obtain the offloading decision of delay-sensitive tasks and the required computing resources.The remaining computing resources are used to calculate the energy consumption sensitive tasks.The simulation results show that the proposed algorithm can reduce the time delay and energy consumption,and can make full use of the limited computing resources to complete more computing tasks according to the requirements of the task.(3)For the edge computing network powered by Wireless Power,a multi-user offloading decision and task scheduling algorithm is proposed by combining Wireless Power Transmission(WPT)technology with MEC technology.In the multi-user binary computing offloading mode,an optimization problem to minimize the maximum delay of system execution is proposed by jointly optimizing wireless power supply time,edge user offloading decision and task scheduling.Due to the optimization problem is a mixed integer non convex optimization problem difficult to solve,we design a two layers of alternate iteration solving scheme,the first layer uses the greedy algorithm and Johnson algorithm based on improved,solving users of wireless power at a given time uninstall decision-making and task scheduling,the second layer based on the method of golden section to get the optimal wireless power of time.Simulation results show that the proposed algorithm has fast convergence speed and effectively reduces the execution delay of the whole system.