Research On Content Caching Optimization In Mobile Edge Computing System

In recent years,with the rapid development of mobile Internet,Internet of Things and the emergence of various new applications,such as virtual reality(VR),augmented reality(AR)and video conference,etc.,mobile users have higher requirements on network service quality(QoS).In order to effectively solve the mobile Internet development challenges such as high-load and low latency requirements,Mobile Edge Computing(MEC)technologies have emerged.MEC improves the conventional wireless access network with service localization and being close to the services deployment,which can provide high-bandwidth and low-latency for transmission capacity,meanwhile effectively reduces the network load and improves the quality of network service.In this thesis,under the framework of MEC cooperative caching,a joint caching optimization mechanism for computation offloading and data caching is put forward,which utilizes the collaboration between edge servers to realize content sharing.Based on theoretical analysis of queuing theory,an optimization model is established to minimize the total delay of the user to obtain the requested content,and an improved genetic algorithm is proposed to solve this problem.Finally,the effectiveness of this mechanism is verified by simulation analysis.Furthermore,in order to further improve the efficiency of edge caching,this thesis proposes a joint optimization mechanism for content popularity prediction and data caching based on machine learning approaches via the framework of MEC collaborative caching and edge controller.Firstly,by analyzing the evolution mode and characteristics of the video popularity,GBDT(Gradient Boosting Decision Tree)algorithm is selected as the prediction model.Then,based on the predicted results,an optimization problem is defined as the objective function to minimize the transmission delay of the requested content,and a heuristic algorithm is designed to deploy the caching optimization scheme.Finally,simulation results vertify that this scheme has a higher caching hit ratio and effectively reduces transmission delay.