On The Game Optimization For Task Scheduling Strategy With Information-visible MEC System

With the coming of the era of the fifth generation(5G)mobile communication technology,Mobile Edge Computing(MEC)is becoming a hot topic in academia and industry,among which the formulation of scheduling strategy is a very challenging problem.In order to optimize the overall performance of the MEC system,the game optimization problem of task scheduling strategy in the MEC system is studied based on the observable information.Firstly,for real-time tasks,a multi-virtual machine MEC architecture is proposed,in which the number of virtual machines is fixed and always active,to ensure the task response performance.The multi-virtual machine queuing model is composed of an M/M/1 queuing model which provides local service,an M/M/1 queuing model which provides transport service and an M/M/c queuing model which provides MEC service.Secondly,for non-real-time tasks,a sleep mechanism is introduced into the MEC server,and a multi-virtual machine MEC architecture based on synchronous sleep is proposed to solve the problem of energy consumption caused by idle resources of the virtual machine.The vacation queuing model is composed of an M/M/1 queuing model which provides local service,an M/M/1 queuing model which provides transmission service and an M/M/c queuing model with multiple vacations which provides MEC service.Then,from the individual and system perspectives respectively,the individual optimal task scheduling strategy and the social optimal task scheduling strategy are proposed.Combined with local equipment information and server state information,the expected sojourn time of the task is solved by dichotomy method,and the individual optimal task scheduling strategy is studied.By constructing Markov chain,the steady state analysis result and system performance index are given.Combining different indexes,the social benefit function is constructed and the optimal task scheduling strategy is studied.Finally,through numerical experiments,the thresholds of individual optimal task scheduling strategy and social optimal task scheduling strategy are analyzed and compared.Under different system parameters,individual optimal task scheduling strategy and social optimal task scheduling strategy are obtained.The pricing scheme of extra sojourn cost in MEC system is formulated,the arrival behavior of tasks is adjusted,and the unification of individual optimal task scheduling strategy and social optimal task scheduling strategy is realized.