Research On Task Offloading Mechanism And Resource Scheduling Strategies In Mobile Edge Computing

As an extension of cloud computing to edge network,mobile edge computing(MEC)has become one of the driving forces of network transformation.By sinking computing power into distributed base stations,MEC provides cloud computing power to mobile terminals with limited computing power and battery life closer,and meets their different resource-intensive and time-sensitive application service requests.Task offloading is a common means for mobile devices to enhance task processing capacity,and reduce time and energy cost by using edge cloud computing power.Reasonable deployment of virtual machine replica copies(VRC)of different applications in the edge network of multiple MEC servers is an effective way for MEC to provide different application resource services for mobile terminals.In the existing researches on MEC,offloading strategy mainly focus on reducing device consumption and system cost by considering the computing resources to make offloading decision.While there are few works to optimize the system total utility by considering the allocation of wireless communication resources.Therefore,it is of great theoretical significance for the practical development of mobile edge computing to design a task offloading strategy that maximizes the system efficiency by joint considering wireless and computing resources.The existing application placement strategies in the edge network mostly focus on the placement of applications' virtual machine resources inside the mobile edge computing platform.There are few works to consider the deployment of applications' virtual machine copy resources on the MEC servers of base stations,which is closer to users,and consider the distribution of user requests.Therefore,it is important to design a multi-application multi-VRC allocation algorithm to deploy VRC in the edge network reasonably.The paper mainly carries out research work form the following aspects:1.To solve the task offloading problem in MEC,user offloading decision and wireless resource allocation problem in the MIMO-OFDM system model are discussed: The interference is generated when edge users of different base stations choose the same wireless resource,and it will affect edge users' uplink data rate.Hence,(1)first,users are divided into the center users and the edge users,then their communication models are respectively discussed.(2)Improved pricing game-based offloading decision algorithm(PGODA)isproposed to get the offloading decision and the wireless resource number of each user,considering heterogeneous nature of different users by introducing the pricing mechanism.Simulation results show that the model can achieve a balance in the end and obtain the optimal offloading decision.Compared with simple game theory,the users in our algorithm behave more “socially”.(3)In the process of wireless resource allocation,in order to prevent edge users of different base stations from choosing the same physical resource block to generate interference,inter-cell interference-avoid communication resource allocation algorithm(IACRA)is proposed.Numerical results demonstrate that IACRA can achieve efficient wireless resource allocation performance.2.To solve the problem of resource deployment of applications in the edge network,the data traffic problem caused by deployment of applications' VRCs in the edge network is analyzed,and we adopt a two-stage multi-application multi-VRC allocation process:(1)Considering the distribution of user requests,and the matching between different applications and MEC servers' capacity,the maximization of high-matching VRC block algorithm(M-HMVBA)for VRC block combination is proposed.(2)Next,the combined VRC blocks will be deployed in MEC servers in edge networks.The minimization average data traffic swap matching algorithm(M-DTSMA)for VRC Block Allocation is proposed.Experiments show that our algorithms can reasonably allocate VRC to reduce the data traffic and provide effective support for the provision of applications services in the edge network.The research achievements of this paper can provide references for task offloading and application resources scheduling in mobile edge computing.