Research On Collaborative Scheduling Algorithms In Multi-access Edge Computing

The emergence of Multi-Access Edge Computing has successfully solved the bottleneck problem of traditional cloud computing networks.Because mobile edges(such as base stations)and mobile devices have specific data processing capabilities,it is not necessary to offload all tasks to the cloud for processing.Therefore,it is very important to determine the optimal task assignment in the MEC system,and a series of algorithms are proposed.However,the existing algorithms ignore the data distribution in the process of task assignment,which limits their application scope.Taking into account the importance of data sharing in the MEC system,this paper studies in detail the problem of ordered task assignment,edge pricing and resource allocation in data sharing Multi-Access Edge Computing system.Firstly,the problem of ordered task assignment is studied in this paper.Many current studies tend to focus on disordered tasks,while ignoring the order relationship between tasks.In this paper,the definition of ordered task is given,and the computing model and the transmission model are proposed based on the MEC system of data sharing.On this basis,the assignment problem of ordered task is discussed,and it is formalized into the mixed integer nonlinear programming problem,which is NP-hard.Then a heuristic algorithm based on clustering is proposed to solve the ordered task assignment problem of Multi-Access Edge Systems based on data sharing.Finally,the correctness and complexity of the proposed algorithm are analyzed.Secondly,the problem of edge pricing and resource allocation in MEC system is studied.A cost model is added on the basis of the ordered task assignment model to model the minimization of user cost and the maximization of marginal benefits.In order to solve the problem of edge pricing,two approximate algorithms are presented to solve the problem of user minimization cost,which are the UTA-G algorithm based on greedy strategy and the UTA-LP algorithm based on linear programming.Then,based on the user allocation algorithm,a reverse induction method is used to price the execution tasks.In this paper,the edge pricing and resource allocation schemes are based on the user dimension and the task dimension,and the EPRA-U algorithm based on user granularity and the EPRA-T algorithm based on task granularity are designed.The complexity of the four proposed algorithms is analyzed.Finally,the performance of the proposed algorithm is evaluated.Firstly,a large number of simulation experiments prove that the HOTA-C algorithm has excellent performance,and the proposed algorithm has high performance in the aspects of delay,satisfaction rate and energy consumption.Then we do a lot of experimental research on the four algorithms proposed in edge pricing and resource allocation.Experimental results show that the proposed user task allocation algorithm can effectively reduce the user's task execution cost,and the proposed edge pricing and resource allocation algorithms can improve the edge revenue and have higher performance.