Research On QoS And Personalized QoE Optimization Method For Mobile Edge Computing Resource Allocation

Mobile edge computing solves both the lack of processing and storage capacity of the client and the problem that the cloud is far away and overloaded.However,the servers on mobile edge computing are small and scattered with limited resources.As the number of Io T devices increases dramatically,it will inevitably lead to a degradation of overall performance and a bad experience for users.In a human-centric time,improving quality of service(QoS)and quality of experience(QoE)becomes important to attract and retain clients,and is further significant for promoting the widespread use of mobile edge computing.In this thesis,we use resource allocation as a means to provide theoretical and methodological support for improving QoS and personalized QoE of mobile edge computing,starting from QoS and personalized QoE of mobile edge computing and proposing different solutions for different problems,as follows:(1)In terms of research on QoS,we propose a QoS scheme to optimize the QoS of mobile devices through radio resource allocation.The scheme predicts the mobile location of Io T devices by Kalman filtering,establishes a mathematical model to estimate the communication reliability,carries out quantitative analysis of performance indicators such as block error ratio,delay,and energy consumption,and proposes an optimization model and method for radio resource allocation.According to the radio resource allocation scheme for mobile devices proposed in this thesis,simulation experiments are carried out based on a real dataset to verify the effectiveness of the scheme,and the experimental results show that the scheme can ensure the connection of mobile Io T devices to the base station while reducing the delay.(2)In the study of personalized QoE,we design a caching policy to optimize global personalized QoE through resource allocation.We use clustering analysis of unquantifiable user and environment information and fit personalized QoE curves based on quantifiable relevant variables.We reasonably design video caching and processing on mobile edge computing and allocate computational resources,storage resources,and bandwidth on mobile edge computing.We optimize the global QoE using the greedy algorithm and projected gradient descent with the momentum method to transform the non-convex problem with constraints into multiple sub-problems.According to the scheme proposed in this thesis,we take simulation experiments based on real datasets.We verify the necessity of personalized QoE and the hit rate of video caching at mobile edge computing.Besides,we get the mitigation effect on the cases of constrained computational capacity,cache space,bandwidth resources,and high load of cloud downlink traffic,meanwhile ensuring the optimization of global average QoE.