Research On Task Offloading Strategy Of Mobile Edge Computing Under Resource Constraints

With the rapid development of the Internet,mobile terminals have become an indispensable part of life and work.Edge computing has been widely studied as a technical method that can process tasks near users.However,in the process of offloading mobile device tasks to edge nodes,there is an imbalance between supply and demand between ESP resources and user needs.Inadequate utilization of computing resources will result in low overall resource utilization,resulting in waste of computing resources and increased operating costs.And with the increasingly complex application scenarios,the workflow composed of tasks is increasingly complex,and there is heterogeneity between different task flows.At the same time,the time complexity of scheduling workflow using the heuristic algorithm is high,and there is a problem of long decision-making time.In view of the current situation that edge devices have weak computing power and are not suitable for dealing with complex problems,there is an urgent need for a computing offload strategy that can make quick decisions and meet resource constraints to solve the optimization problem of computing offload under mobile device resource constraints.Therefore,the main research contents of this thesis are as follows:(1)Aiming at the imbalance between the supply and demand of computing resources,a single-round double auction utility optimization algorithm(SRDAUO)is proposed.First,SRDAUO obtains User-VM matching.Secondly,use Incomplete Information Game to construct bidding strategy to improve system utility(system profit and resource utilization rate);Finally,SRDAUO obtains the computing offload scheme through the auction algorithm.The computing offload scheme obtained by this algorithm performs well in a resource-constrained environment;In addition,SRDAUO has also achieved a high success rate and user profit in a non-resource-constrained environment.It can complete the calculation unloading task well.(2)To solve the problem of computing offload in the context of user mobility and energy consumption constraints,the Multi-User Energy Constraint Time Optimization Algorithm(MU-ECTOA)is proposed.MU-ECTOA includes three stages: cluster analysis,evaluation of performance indicators and determination of unloading objectives.In the first stage,it is classified according to the characteristics of workflow tasks;In the second stage,the unloading plan is planned,and the completion time of different unloading plans is evaluated.In the third stage,the optimal unloading plan is selected for unloading and the actual completion time is obtained.Through experiments,it is concluded that the task scheduling and computing offloading algorithms proposed in this thesis have certain advantages over the comparison algorithms in terms of delay,user profit and system utility.The latter has certain advantages in delay,energy consumption,algorithm reliability,algorithm time complexity,load balancing,etc.