Task Offloading Strategy With Clustered Virtual Machine And Sleep Mechanism In Mobile Edge Computing

With the emergence of a large number of energy-intensive and computationally intensive applications,low-latency and low-energy task processing requirements pose severe challenges to mobile edge computing(MEC).The paper takes into account user response performance and system energy-saving level,introduces periodic sleep mode and half-period sleep mode on the MEC side,integrates virtual machine clustering technology,and studies the MEC task offloading strategy and system performance.Firstly,in order to improve the energy-saving level of the system,a MEC task offloading strategy that combines virtual machine clustering and periodic dormancy is proposed.Under the control of the sleep timer,the standby module sleeps periodically until a task arrives,and the standby module virtual machine returns to the active state in due course.Establish a synchronous multiple vacation queuing model for some service desks,construct a three-dimensional continuous-time Markov Chain(CTMC),use the quasi-birth-death process and matrix geometric solution method to give the average task delay and system energy-saving rate expression.Secondly,in order to improve the user’s response performance while guaranteeing the energy-saving level of the system,a MEC task offloading strategy that integrates virtual machine clustering and half-period sleep is proposed.If a task arrives before the standby module sleep count reaches the threshold,the virtual machine enters the active state at the same time after the sleep timer expires;if no task arrives after the standby module sleep count reaches the threshold,the virtual machine still enters the active state at the same time.The system is modeled as part of the service desk synchronization multi-level adaptive vacation system,a four-dimensional continuous time Markov chain is constructed,and the expressions of the average task delay and the system energy saving rate are derived.Thirdly,in order to verify the validity of the MEC task unload strategy,in view of the synchronous multiple vacation queue model with synchronous multi-stage adaptive vacation queue model,based on the theoretical analysis of the numerical experiment,the working mechanism simulation experiment based on the strategy of task unloading,reveal the local distribution probability under different task arrival rate of task average latency and energy saving rate of the system.Finally,in order to balance the relationship between the average task delay and the system energy saving rate,the system cost function was constructed.For the above two different MEC task unloading strategies,a "teach-and-learn" method was introduced to improve Pigeon-Inspired Optimization(PIO)to achieve the goal of minimizing the system cost.The optimization scheme of MEC task unloading strategy is given.