Mobility-Aware Cloud-Edge-End Collaborative Application Offloading Strategy

With the rapid development of integrated circuit and information technology,the number of mobile terminal equipment presents a blowout growth trend.Subsequently,a large number of computation-intensive applications emerged,such as virtual reality,augmented reality,face recognition and so on.As a result,cloud computing arises at the historic moment.It solves the limitation of terminal equipment capacity by virtue of its powerful resource,but the delay of task transmission is large.Edge computing has also been widely studied and applied because of its low transmission delay,but its computing benefits are very small when dealing with applications with large resource requirements.Therefore,this thesis proposes a mobility-aware cloud-edge-end collaborative application offloading strategy,the strategy fully combines the huge advantages of abundant resources of cloud computing and low transmission delay of edge computing,and takes into account the mobility of users to make reasonable offloading choices and shorten the completion time of computing tasks.The main research contents of this thesis are as follows:(1)To solve the problem of user movement in cloud-edge-end collaborative computing,this thesis proposes a generative adversarial network model based on attention mechanism for pedestrian trajectory prediction.Firstly,four key feature information of pedestrian interaction is modeled,and the influence of other pedestrians is quantified by the attention mechanism.Through the minimax game between generator and discriminator,the accuracy of model prediction is further improved.The experimental results show that the ADE and FDE indexes of this method are optimal,and the accuracy of prediction is significantly improved.(2)To solve the offloading problem of computing tasks,this thesis proposes a mobility-aware cloud-edge-end collaborative application offloading algorithm.Firstly,the algorithm actively caches the results of computing tasks on the MEC server and the cloud.On this basis,it makes task prediction based on the linear regression model to get the offloading time of different offloading nodes.Then the pedestrian trajectory prediction model is used to predict the user’s moving trajectory,and the user’s stay time within the range of each offloading node is obtained.Finally,the offloading destination of the computing task is determined.The experimental results show that the algorithm proposed increases the offloading success rate of tasks compared with some traditional offloading algorithms and greatly reduces the completion time of computing tasks.(3)Based on the above mobility-aware cloud-edge-end collaborative application offloading algorithm,this thesis designs and realizes a mobility-aware cloud-edge-end collaborative computing verification system,and use three kinds of computing applications with different characteristics to test whether the verification system can successfully completes the calculation offloading and shorten the completion time.Test results show that the system can fully realize the expected function.