Research On Mobile User’s Service Migration And Offloading Strategy Based On Three-way Decisions In Edge Computing

With the rapid development of communication technology and Internet of Things,the number of user terminal equipment has increased rapidly.Meantime,computationally intensive applications such as virtual reality and augmented reality have emerged.Resource-constrained terminal devices are unable to meet the requirements of such applications for real-time performance and computing capacity.Therefore,terminal devices need to seek third-party device service support to process such applications.Mobile edge computing emerged.In the mobile edge computing environment,devices can get service support by connecting to edge servers.However,due to the limited location and service range of edge server,edge server can only provide service support to users within its service range.In the real pedestrian movement scenario,the user’s movement is random,and the terminal device are always with the user.Due to the limitation of edge server’s service range and the randomness of terminal user’s movement,it is difficult to provide users with efficient computing service and stable service quality at the same time.Service migration has become an important issue in mobile edge computing.The current service migration strategy based on the two-way decisions makes the decision of whether to migrate or not to migrate when the user’s accurate mobile information is not known.The wrong decision may lead to multiple service migration or even service interruption,which seriously damages user experience.Therefore,in the mobile edge computing environment,we first propose a service migration strategy based on three-way decisions.Secondly,for users who do not need service migration,a computational unloading strategy based on genetic algorithm and sequential three-way decision is proposed to ensure timely completion of high real-time service.The main research contents of this dissertation are as follows:(1)In this dissertation,the user’s movement tendency is defined according to the changes of distance relative to the edge server in a period of time,and the evaluation function is defined according to the user’s movement tendency and three-way decisions in multiple periods of time.Finally,the evaluation function of each user is calculated,and users are divided into migration region,delay region and non-migration region respectively according to a pair of given thresholds and .(2)For the migration region,we propose a migration strategy based on migration energy consumption and service delay time.We define the latest migration radius and the lowest energy consumption migration radius respectively according to user location and the size of services.Service migration based on this strategy can effectively reduce migration energy consumption cost and service delay time,and improve user’s experience of service.For the delay region,edge server continues to collect user movement information for further service migration decision,so as to improve the decision accuracy.(3)For the non-migration region,we propose a service computing offloading strategy based on three-way decisions and genetic algorithm to further reduce service delay and optimize service experience.First,we build a task partitioning model of equivalence class,offload tasks to the best platform,and initialize task sequence.Then,we introduce the idea of sequential three-way decisions,and put the bad task sequence into the rejection region to reduce the search space and speed up the algorithm execution.In addition,in the algorithm,we propose the overlapping task method of crossover process,which can generate and maintain the better solution sequence into the next generation.