Task Offload Strategy Of Mobile Edge Computing Based On Lyapunov Optimization

With the fast development of mobile devices,related smart applications have significantly changed people's lifestyles.However,because of the limited computing and storage resources of mobile devices,Besides,the task offloading methods in traditional cloud computing,which have high latency,network congestion,and long transmission distance,These traditional methods are unable to meet the computational requirements of computationally intensive and delay-sensitive tasks.So the Mobile Edge Computing(MEC)technology becomes popular,it sinks cloud computing and storage capabilities to the user's near-end network nodes,and can also provide computing services while meeting low latency requirements.Therefore,offloading tasks to the MEC server for calculation can improve task calculation efficiency,reduce task delay,reduce mobile device energy consumption and task offload costs.There are many factors affect the task offloading in MEC,such as energy consumption,delay,and cost of task offloading.Therefore,how to efficiently offload tasks is an important problem that needs to be solved urgently.Until now,many scholars have studied the problems of minimizing the cost of unloading tasks,minimizing energy consumption and delay in MEC,but there are some insufficiencies:(1)In the study based on minimizing the cost of unloading,task unloading in MEC is not considered Optimization of the cost(including channel transmission cost and task calculation cost),and the possibility that other idle mobile devices can provide computing services;(2)Based on the study of minimizing energy consumption and delay,no consideration is given to ensuring the stability of the queue at the same time Reducing task delay and mobile device energy consumption,and the existing research on reducing mobile device energy consumption and task delay,the method used depends on the past state of the system parameters,cannot make dynamic offload decision every time slot.Because of the above deficiencies,this article conducts the following studies:(1)When multiple mobile devices generate task offload requests at the same moment,we want to achieve the goal of reducing task offload costs.On the premise of ensuring queue stability,this paper proposes a dynamic offload algorithm based on Lyapunov optimization.According to the task offloading situation,we establish a task queue model,at the same time,we also propose a task offloading strategy based on task offloading cost and delay balance.In each time slot,the mobile devices make a dynamic offload decision based on the generation of mobile device tasks.According to the dynamic offload decision,the mobile devices determine where is the best place for the tasks to be calculated.For example,tasks are calculated locally or allocated to other mobile devices with free computing resources for calculation or allocated to the MEC server for calculation.In the end,the simulation shows that the algorithm can reduce the cost of task offload and also ensure the stability of the queue.(2)When multiple mobile devices' task is about to be offloaded dynamically,in order to achieve the goal of minimizing mobile device energy consumption and task delay,we establish the task cache queue and MEC server queue.Based on the Lyapunov optimization method,we obtain the optimal CPU frequency of the mobile device when the task was locally computed,we also solve the optimal match of the dynamic unloading of the task.In each time slot,the MEC system makes a dynamic offload decision depend on the queue parameters,that is,the task either is calculated locally or offloaded to the MEC server for calculation.In the end,the simulation shows that the algorithm can effectively reduce the mobile devices energy consumption and task delay.