Research On Computation Offloading And Resource Management Mechanism Based On Heterogeneous Resource Cooperation In Edge Networks

With the rapid development of mobile internet technology,smart phones,tablets,and other mobile devices have become an indispensable part of people’s daily life.However,the computing resources and battery capacity of these devices are usually limited,resulting in difficulties to effectively support computation-intensive applications.By deploying edge servers near base stations,Mobile Edge Computing(MEC)pushes lightweight cloud computing platforms to the edge of the network,offloads computing tasks of mobile devices to edge servers,and efficiently manages various resources in the network to meet user needs and improve resource utilization efficiency.At present,MEC has attracted extensive attention from the academic community,but it also faces significant challenges brought by the diversification of computing services and the short-term centralization of service requests.On the one hand,computing offloading needs to deploy a specific service environment on the server to adapt to different computing services,and users’ various quality requirements for services increase the diversity of service environments.On the other hand,typical network scenarios deploy personalized communication and computing resources to provide services,and the aggregation of user requests in a short period brings resource scheduling strain among scenarios.To solve the above problems,this thesis studies the computation offloading,and resource management mechanism based on heterogeneous resource cooperation.The details are introduced as follows:Aiming at the difference of service environment caused by the diversification of computing service requirements,this thesis proposes a computation offloading mechanism based on quality level awareness model,which jointly optimizes equipment energy consumption and service delay.Firstly,considering the number,dependency and quality level of computing tasks,a quality level awareness model is proposed to realize the unified representation of computing services.Secondly,the computation offloading problem under quality level constraints is formulated,and the optimization objective is to minimize energy consumption and delay.Then,a pre-scheduling algorithm of computing tasks is designed to realize parallel processing of tasks and accelerate the convergence of the algorithm.On this basis,an offloading decision algorithm based on quality level awareness is designed to obtain the offloading strategy of computing tasks in polynomial time complexity.Simulation results show that compared with the existing algorithms,the proposed mechanism optimizes the energy and delay efficiency,and improves the convergence performance of the algorithm.Aiming at the resource scheduling problem caused by short-term centralization of service requests,this thesis proposes a resource management mechanism based on cost-aware model,which can improve the utilization efficiency of resources in the network under the premise of satisfying the delay constraint of users.Firstly,for the coexistence and overlapping coverage of multiple network scenarios,considering the predeployed resources in the scenario and the usage status of computing and communication resources between scenarios,a cost-aware model is proposed to realize a unified representation of resource utilization.Secondly,the resource management problem under delay constraint is formulated,and the optimization objective is to minimize the resource scheduling cost.Then,a cost-aware resource allocation algorithm is designed to jointly allocate computing and communication resources in the network.To screen the feasible solutions of the resource allocation algorithm,the probability constraint tolerance algorithm is designed.Simulation results show that compared with the existing algorithms,the proposed mechanism reduces the overall resource management cost of the system and the resource usage overhead of a single user.