Research On Task Offloading Mechanism Based On Energy Optimization Of Mobile User In Cloudlet Environment

With the blowout of mobile application data,the problem of energy consumption for mobile devices is becoming more and more prominent.It can save energy effectively by offloading the tasks of mobile device to the nearby cloudlets.However,in real application scenarios,the connection between cloudlets and users is discontinuous.In the study of discontinuous connection,only 10% of energy consumption can be saved when the task is offloaded to the cloudlet compared to local execution,which indicates that there is still a higher energy saving space in the real task offloading scenario considering discontinuous connections.Therefore,based on the research of real task offloading scene in the cloudlet environment,it is of great theoretical and practical significance to propose energy-efficient task offloading algorithm to save the energy consumption of mobile devices.The energy-efficient task offloading algorithm studies the discontinuous connectivity caused by the randomness of user mobility,cloudlet distribution randomness and resource limitations,and is designed to minimize the energy of mobile device before the deadline of tasks.First,based on the limitation of cloudlet resources,a cloudlet access mechanism based on semi-Markov process is proposed,which enables users to maximize the use of cloudlet's resources.Then,based on the randomness of user mobility,the success probability of task offloading is derived.Then,based on the global prediction,the number of tasks that can be offloaded to cloudlets in the process of user movement is calculated,and the optimal CPU frequency of the mobile device is further deduced.Finally,based on the above conclusions,a dynamic task offloading algorithm—ETOA based on probability is proposed.In the process of user mobility,the goal of minimizing the energy consumption of mobile devices is achieved by comparing the probability of offloading successfully probability and the threshold probability of the task.The test results show that the cloudlet access mechanism based on the semi-Markov process makes the resource utilization of the cloudlet remain above 93%,which meets the resource requirements of the mobile user.And the experimental data also shows that the task offloading algorithm--ETOA can save energy effectively on the basis of ensuring the task completion rate,and can save up to 50.5% of the energy consumption compared with other algorithms.